ASSISTANT secmaav or oerewse WAWNGTON D C 10301 June 26 l978 r3 Moan ANALYSIS WW awocvawa rmu EL Ii ll MEMORANDUM FOR THE SECRETARY or DEFENSE SUBJECT Interim Report on the po-Ia Study Entitiedu'nodernizi tl i of the ICBM Force ACTION MEMORANDUM Attached at Tab A is the remaining portion of the Interim Report on the 90-18 study entitled Modernization of the ICBM Force As you will recall a couple of weeks ago I sent you the first two papers prepared for this report These papers analyzed the options of deploying MX in fixed silos and launch-under-attack An appendix on Minuteman surviv- ability was also included - Our purpose in sending you these papers was to seek your agreement that these Options should not be pursued further in the evaluation of the retaliatory effectiveness of alternative forces It was not our intention to drop these Options from the study but to discuss them only in the context of perceptions including effects on strategic indices and programmatic considerations Although I believe you agreed with us on the limited military effectiveness of MX in silos in response to your comments we will keep this option in the analysis and evaluate its effectiveness with and without launch-under-attack For the Interim Report we have simply deleted from the paper on MX in silos the section that asked for your approval to handle this option as indicated above The remainder of the paper remains essentially unchanged so I am not sending it to you again I3 document al- E's We are not sure we have adequately responded to your comments on the Launch-Under-Attack paperz Your times are not right EA messages will be pre-cut DSP follow-on can begin to give assessment The Air Force has reviewed the paper and provided some changes incor- porating SIOP-SB timin At least one of these chan as is 51 nficant i in the case where the DECLASSIFIEI UNDER AUTHORITY OF THE SECURITY CLASSIFICATION APPEALS PANEL 13 0 13526 SECTION ISCAP No We have also added a sect on on ncrease capability from DSP follow-on The revised version of the paper is at Tab 8 with the changed numbers and sections underlined in red If you still have problems with the paper we would appreciate your guidance so we can properly address this issue in the effectiveness analysis mw w 13526 section ll Inna Jane ri Upon removal or attaciunants been-us At Tab A you will find the other papers in the Interim Report A These include - The Introduction and Sumnary section which briefly sunlnarizes each of the papers - a paper entitled The Strategic Options Matrix and - an appendix entitled Analytical Methodology - One appendix The Threat is not included because of the codeword classification it is available if you would like to see it In the paper on the strategic options matrix we are seeking your approval of our revised matrix for generating force mix alternatives and our proposed approach to conducting the cost-effectiveness analysis The methodology appendix documents our proposed approach to strategic force mix analysis which I believe represents a noteworthy accomplish- ment of the study effort to date Taken together these papers describe the work we plan to do this summer to evaluate the cost-effectiveness of alternative force structures Since a lot of manhours will be'involved we should consider now whether this type of information will be helpful to you in making the decisions we face in planning our future strategic force posture Although a large number of comments received from 0008 the Services BIA and other OSD elements have been incorporated into this report it is not being forwarded as a coordinated product As we indicated at the beginning of the study our intention is to include with the final report all comments not incorporated in the study itself By separate cover we are forwarding to the Executive Study Committee the entire Interim Report including the first two papers modified in response to your comments Four more papers will be forwarded to you around the end of July These papers are on the Tried and hedging against uncertainty 2 quick reSponse hard-target kill capability multiple aimpoint systems and 4 a treatment of the perceptions and programmatic aspects of basing ax in existing Minuteman silos These papers are currently being written and reviewed None require decisions they are being prepared to highlight important issues relevant to the final report in September Russell Murray 2nd Assistant Secretary of Defense Program Analysis a Evaluation Attachments cc DepSecDef ll I - Prepared hy copies to Special Assistant Military Assistant OSD Maiiroom msn me Bends m OASD me A INTRODUCTION AND SUMMARY A Introduction In response to the NSC Directive of August 24 1977 subject Follow-On Studies for the Secretary of Defense initiated a non study to address the following related issues modernization of the ICBM force 2 the future of the Triad concept and 3 potential requirements for hard-target kill capability This interim report on that study consists of the following discussion papers 1 MX in Existing Minuteman silos 2 Launch-Under-Attack and 3 The Strategic Options Matrix The papers are briefly summarized in the next sub-section Also included in this report are three appendices l Appendix A Minuteman Survivability is the executive summary of a recent SALT interagency paper on Minuteman survivability - The bottom-line conclusions are the Soviets will be capable'of reducing Minuteman silo survivability to very low levels by the early 19805 and they should have confidence in their ability to do so Although they may have major unquantifiable uncertainties such as fear of U S launch-under-attack tactics -technical uncertainties associated with- reliability accuracy etc should not be very great 2 Appendix 8 Analytical Methodology describes in detail the analytical methodology to be used in this study to evaluate the retaliatory capability of alternative U S force postures The methodology has been accepted by the study working group and will enable us to evaluate relative force effectiveness over a broad range of targeting objectives under fairly realistic constraints 3 Appendix c Threat is the description of the threat to be used in the study It was prepared by BIA and will not be circulated unless requested because of its classification level There are four other papers currently in preparation which will be available at the end of July i ieferred to as the HxiTriad study throughout this report or 0 The Triad and Hedging AgainSt Uncertainty - Programmatic and Perceptual pros and cons of silonbasing - Requirements for quick-response hard-target kill capability and - Multiple Aimpoint System issues 8 Summary Each of the three discussion papers is summarized below T Mx in Existing Minuteman Silos This paper deals only with the added retaliatory capability Of silo based MX Due to the amount of recent high-level interest a separate sub committee of the MX Triad study working group is being formed to write a paper on the programmatic and perceptual pros and cons of this option The lessons learned from the SALT interagency study on Minuteman survivability lead to the following conclusions Even if MX were deployed in the hardest of current 51105 and even if the Soviets did not adopt special tactics against the silos containing Mx HX survivability would almost certainly be very low 2-9 10% for a media flgure - 123 0 13526 section The Soviets could reduce nx survivabilit to even lower levels by referentially targeting and or by using more than This would be so even if there were substantial fratric e The Soviets could have high confidence in their ability to reduce x to very low levels there is no evidence to support the notion that Soviet calculations would differ greatly from our own at least after about l983 The paper goes on to quantitatively refute the commonly made argument that surviving MX in silos would be significantly more effective that an equal number of MM 11 survivors If the U S chooses to ride out an attack there are two options I prevplanned targeting with redundant coverage to account for low PLS pre-launch survivability and 2 optimal post-attack retargeting The calculations show an almost imperceptible difference in retaliatory capability between 200 silo-based MX and MM if the other legs of the Triad work as anticipated even with optimal post attack retargeting Even if one of the other legs fails catastrophically the incremental capability added by the few MX survivors is small 3 This analysis is intended to counter the connmn argument that because 100 MM could destroy roughly 50% of Soviet industry the greater capability of a few surviving MX is significant The argument is true only if the prime urban industrial targets were left unattacked as a result of a catastrophic failure of both and bombers 2 Launch-Under-Attack This paper seeks to draw a clear distinction between a leunch-under attack LUA capability to enhance deterrence and LUA as a high confidence tactic to maintain an effective ICBM force if deterrence fails A credible LUA capability may enhance deterrence may serve the domestic purpose of dampening concerns about declining Minuteman survivability and may be cheap relative to mobile MX or Trident 11 However the paper argues that an operational LUA capability is not a viable alternative for modernizing the ICBM force primarily because 0 The time available to analyze attack assessment infornetion and execute a res onse prior to the arrival of Soviet ICBMs would be do best and - if the Soviets also attacked our 31 0 13526 section To have confidence in this tactic would require a system for transferring authority to a properly briefed and survivable figure - quickly in the event communications with the President failed Further should the President die in an attack timely and proper transfer of authority over nuclear release must be ensured Both requirements raise some fundamental constitutional questions The President and his potential successors would have to be thoroughly familiar with the attack assessment systems and be prepared to make a decision in very little time Otherwise we would be accepting an unplanned Dyad with a significant loss in our retaliatory capability 0 Finally although the President could decide to retaliate in a few minutes he shouldn't as a matter of policy have to do so 3 The Strategic Options Matrix The first objective of this paper is to seek Secretary of Defense approval for our proposed change to the strategic options matrix presented in the study outline In that outline we addressed the queStion of the future role of the ICBM force in terms of the relative role of the ICBM force We believe the statement that best reflects the immediate issue for decision is not the ultimate deployment level but rather Do we want an effective Triad do we want to modernize the ICBM force so that all three legs of the Triad make a significant contribution to the retaliatory capability of our strategic forces The refinement of the ultimate size of the ICBM force can be left to be worked out after SALT is settled and the review of our targeting objectives has been completed Acceptance of this change leads to the revised matrix shown below y Revised Strategic Options Matrix Do We want An Triad with Modernized Effective Triad ICBM Force Survivable Hard Target capability Unbalanced Triad With No ICBH Modernization Extensive Quick- Response Capability Extensive Non-Quick- Response Capability No First Strike Capability Against Silos Limited Counter- Silo Capability The second objective is to seek approval of our proposed approach to conducting the cost-effectiveness analysis of the alterna- tive force structures that are developed to fit the revised Options matrix We argue in the paper that it is not appropriate to study either equal effectiveness or equal cost forces Accordingly we recommend that alternative forces be studied that are programmatically feasible and consistent with SALT although not necessarily such as to fill out the SNDV aggregate -- this being more nearly the type of decision facing the Administration budgetary etc are also proposed was A number of excursions threat A Introduction The Secretary of Defense and other high-ranking Administration officials have publicly discussed a U S laUnchaunder-attack LUA option to demonstrate the great uncertainty facing a Soviet leader contemplating a first strike against the U S Such statements may enhance deterrence jj_they are credible they any also serve the domestic purpose of dampening concerns about declining Minuteman survivability In discussing LUA we must draw a distinction between LUA to enhance deterrence and LUA as a tactic to maintain an effective ICBM force if deterrence fails There is a danger that we may come to embrace LUA as a substitute for deploying a more survivable system without fully recognizing its limitations and implications In other words it's one thing to talk about LUA and quite another to structure our forces such that their effectiveness is dependent on it The following discussion is intended to raise a number of questions and highlight some important problems associated with an operational LUA capability Let's begin our discussion by looking at current U S capa- bilities and limitations for launch-under-attack B Current U S Capabilities and Limitations for Launch-Under-Attack ICBH force has the erational ca-abilit for are controlled 12 0 13526 section 3 3610 5 CUrrent and planned U S missile surveillance sensors are intended primarily for tactical warning The Defense Support Program DSP has recently been modified to include some raid counting capability but it is not capable of high confidence attack characterization For example it cannot indicate that only force is under attack Similarly in the near term the Ballistic Missile Early Warning System BMEWS will be limited to providing further evidence given DSP information of an impending attack on Minuteman Some capability to assess attacks on Minuteman resides in the Safeguard Perimeter Acquisition Radar Attack Characterization System PARCS but data from this sensor is probably not available to the NCA in time to support the execution decision process for most By 1985 we can expect a better capability to characterize the magnitude of the threat to Minuteman and with the BMEWS range resolution upgrade we can expect a capability to indicate classes of targets that are under attack including an indication that the NCA is targeted a -- SUBJECT TO GE EXECUTIVE once a INTERVALS UT 05-1 1 WNG RADED ECLASSIFID COFY mama-9 0F 2 After detennining that the U S ICBM force is under attack there remains a finite amount of time to refine attack assessment and to execute a response prior to the arrival of Soviet ICBM reentry vehicles on Minuteman silos The time available is very scenario dependent as seen in Figure l Today preparation of an EAM regardless of scenario However the time required for transmission of the EAM to first launch of ICBMs varies from available Given no fireout constraints salvo launch of flyout time and availability of the primary landline network the execution message preparation may be initiated as the messa re aration If it is With the current constraints would have to be assumed that an earl 13526 section within the time windows defined by c3 availability there is a continuum of attack assessment data that will be available to the NCA before the EAM decision is required DSP reports of simultaneous Soviet launches would begin to arrive roughly before impact of Soviet depending on their launch point and - before ICBM impact on Minuteman silos assuming a nominal flight time of 30 minutes SLBM radar and BMEWS reports would arrive about I before ICBM impact on silos PARCS data on incoming RVs would arrive for NCA use from Thus i C3 nodes and the NCA are not attacked and if osp SLBM radar and smf s assessment is required in the decision process the decision time available to the NCA is about If PARCS data is required at most-t ree minutes 0 decision time would be available lf_the C3 network is attacked there would be no decision time available to use Eh current attack warning or assessment sistem other than the 13526 section General Slay in his recent testimony before the Senate Armed Services Committee said We do well with DSP BMEHS and the combination to give an indication that Minuteman is under attack To confirm a large raid on Minuteman or to confirm a threatening raid on Minuteman we have no capability with DSP or BHEHS or with the PARCS a system we have inherited from the Army We have some capability but we sa it is poor and that the unreliability of that indication would bet percent Exmxawm 58% me dd - 4 independent of the sensitivities to availability of attack assessment information and C3 assets the critical factor in getting the execution message to the ICBM launch crew prior to the impact of Soviet RVs on the silos is the survivability of the President and the time required for Presidential decision In the absence of strategic warning the time required to contact the President and receive his execution decision may vary from a If during this time either the NCA or the President are attacked by early arriving SLBM warheads the execution decision may c Potential use Improvements 13526 section museum The draft C31 study on Missile Tactical Warning and Attack Assessment assesses current and potential capabilities of U S tactical warning and attack systems including the DSP system to provide timely and credible information to support a launch under attack option for the Minuteman force The study concludes that our current system would provide sufficient information to support a LUA decision if we were willing to launch Minuteman on the basis of data indicating only that CONUS were under attack by a large force If on the other hand more precise attack information were required for an LUA decision determining that 150 450 or had been launched toward central CONUS an improved DSP system would be needed to provide more accurate launch count and trajectory azimuth information In the near-term 1980 completion the study proposes improving DSP azimuth accuracy by upgrading DSP software to permit processing of booster reentry data With this improvement DSP would provide about eight minutes after launch or about six minutes after initial DSP readout high confidence information for determining whether 35-18 or missile trajectories were heading toward the Minuteman complexes In the longer-tenn mid-19805 at the earliest the study proposes improving DSP booster count and azimuth accuracy as part of the Sensor Evolutionary Development program Given a decision to support these improvements now this evolutionary upgrade to the DSP sensor would within 2 minutes of launch permit the system to count boosters within about 10% and provide azimuth accuracies sufficient to determine that a Minuteman threatening attack had been initiated While these improvements would improve the accuracy and detail of the information provided by the DSP system it should be noted that they do not address related issues that also bear on the credibility and feasibility of LUA the possible need for dualwphenomenology the vulnerability of U S sensors and their sugporting communications systems and the vulnerability of the strategic system including the RCA to J stones D Scenario Assumptions and LUA Alternatives To discuss a LUA option relative to other alternatives it is necessary to examine the range of possible scenarios for a strategic exchange and to recognize that having a realistic LUA capability for all or even most scenarios is very different from having such a capability under selected circumstances Specifically if the decision were made to retain a Triad and make either Minuteman or a relatively low cost silo-based HX system effective through LUA we would have to ensure a reliable LUA system for all realistic scenarios including Soviet minimum warning attacks aimed at disrupting such a capability To do otherwise would be to accept under certain circumstances an essentially unplanned catastrophic loss of a Triad leg The resulting Dyed would be considerably less effective than would be the case if we made a decision to rely only on the Dyed force and preplanned its employment accordingly Accepting a Dyad on day-to-day alert and a Triad on generated alert with much greater capability could tend to destabilize the strategic situation in a crisis by making a surprise attack far more profitable 0n the other hand if we were willing to accept a Dyad under worst case assumptions a bolt-out-of-the blue attack but wanted a Triad for what we believed to be a more likely scenario generated alert following a period of tension then we should compare LUA with poten- tially cheaper ICBM modernization options than have heretofore been proposed For all MX systems currently under consideration cost is driven largely by the need to survive any attack It may be possible to develop a lower cost mobile ICBM system that would be survivable under generated alert but not under day-to-day alert A LUA strategy would eliminate some inherent advantages of the ICBM force principally flexibility Further it would give the Soviets control over when our ICBM force was used An LUA decision tree which summarizes the relationship between scenario assumptions and alternatives for an operational LUA capability Can be depicted as shown-in Figure 2 E Requirements and Problems Associated With An Operational LUA Policy Let's look at some of the requirements and problems associated with the LUA policy for all cases versus such a policy under selected cir cumstances imam Figure 2 LUA Decision Tree Do we want an operationei 4 LUA in lieu of a survivable Do we want the LUA capa Do not consider LUA as a viabie aiternative to a Yes to provide a Triad more survivabie repiace- under reasonabie lscenarios ment for Minutemen 1 53iine circumstances under which we wouid be to accept other alternatives providing simiiar a Dyed Ye 2 Compare LUA-with Determine if such a cape- is feasibie Consider as fuii fiedged oiternative for modernizing the ICBM force l LUA As a Viable Alternative Under All Reasonable Scenarios The first requirement would be a greatly improved warning and attack assessment system that the President or his successors could rely on to accurately assess the type and magnitude of the attack These systems are vulnerable today and a fundamental question is Can we deploy a reliable and survivable attack assessment system It would require a President and his successors thoroughly familiar with the attack assessment systems and the execution options as well as prepared to launch the ICBM force in minutes based only on information supplied by available sensors If the President doesn't survive could we get the attack warning and assessment information to one of the President's successors located in some other part of the country or the world and could he or she execute the SIOP in time Would this individual be able to ascertain the nature of the attack on the NCA or even who was responsible or it How would the system work under the following scenario The Soviets launch a coordinated and ICBMs bolt-out-of-the-blue attack on the U S The Soviet objective is to destroy our ability to retaliate to the maximum extent possible so they go after our C3 Suppose further that as art of the attack were fired on the White House during a Cabinet meeting or a reception for some dignitary would the President and the Cabinet survive If not would the next in line wherever he miiht be located at the time 12 0 13526 section Another scenario which is hard to cope with if on LUA is the following The Soviets attack satellites BMEWs and other Warning Centers with non-nuclear munitions perhaps as part of a general nonunuclear attack Do we immediately escalate to the massive use of nuclear weapons obviousli does not solve our irolleml Eu onli increases The purpose here is not to dream up the most ghastly scenario but rather to illustrate the depth of planning and coordination that would be required for an operational LUA strategy It appears that if we desired a LUA strategy that would be effective under most scenarios we could a priori choose one of the following tactics to ensure such a capability Make the LUA response automatic given loss of certain sensors or communications enforce a requirement that a qualified successor to the President be survivable at all times airborne in NEACP except those crises where the President himself would be secure 8 Tactic is probably unacceptable to everyone and therefore can be dropped It would certainly present the danger that we could launch our ICBMs as a result of mechanical failures or the clever work of some terrorist Tactic raises profound questions regarding delegation of authority civilian control and crisis stability Should the President die in an attack can we ensure that his successor will have the required survivable anti-jam C3 to support a LUA strategy and at the same time guarantee that a temporary communications failure etc would not lead to a premature transfer of authority over nuclear forces 2 LUA As a Strategy Given a Crisis or Strategic Alert It is clearly easier to develop a LUA strategy in a generated alert situation than for a bolt-out-of-the-blue attack Given sufficient warning the President or his appointee could be airborne in the NEACP with all systems alert and prepared for any possible attack It is doubtful that the President himself would fly NEACP in anything but the gravest situation which means that even in generated alert provision for transfer of authority would have to be made in a timely manner Never- theless such a system could probably be made to work today although it would still necessitate a highly formalized procedure for declaring a crisis History provides abundant examples of surprise attacks for which strategic warnings were ignored F Tar etin - A related issue concerns how we would target the ICBM force under a LUA policy At the outset we must recognize that LUA only salvages a fraction of the targeting capability which a secure survivable ICBM force would have It seems probable that we would want options to respond to different types of attacks in an appropriate manner while controlling escalation For example suppose the Soviets launched a countersilo-only attack we might want to respond in kind 0n the other hand if the Soviets were attacking both counterforce and countervalue targets we might want our ICBMs to strike at least some countervalue targets as a hedge against problems in the other legs of the Triad Of course we wouldn't want to attack countervalue targets if the Soviets had only gone after our ICBMs It's clear therefore that there would be more to a LUA strategy than just getting a message out to execute the force the message would have to include specifically which missiles against which targets This selective response capability would require excellent attack assessment pro-planning and training by the President and his successors and the ability'to decide and execute the appropriate response in minutes G Summary An operational LUA capability is not a viable alternative to modernizing the ICBM force primarily because This issue is being addressed in detail by a sub-panel of the Targeting Study This is not to say we should not make an LUA option more credible to capitalize on its positive impact on deterrence IE 0 13526 section The time available to analyze attack assessment information and execute a res onse rior to the arrival of Soviet ICBMs would be end in man cases as ecially if the Soviets also attac our To have confidence in this tactic would require a system for transferring authority to a properly briefed and survivable figure quickly in the event communications with the President failed Further should the President die in an attack timely and proper transfer of authority over nuclear release must be ensured Both requirements raise some fundamental constitutional questions 3 The President and his potential successors would have to be thoroughly familiar with the attack assessment systems and be prepared to make a decision in very little time Otherwise we would be accepting an unplanned Dyed with a significant loss in our retaliatory capability 4 Finally although the President could decide to retaliate in a few minutes he shouldn't as a matter of policy have to do so PAPER THE STRATEGIC OPTIONS MATRIX A Introduction The MX Triad study is designed to illuminate two important issues that require Presidential decisions before 000 can effectively structure a future U S strategic force posture They are l What should be the future role of the ICBM force 2 What are U S requirements for a hard-target kill capability Each of these issues will be addressed separately in this study although as will be seen alternative force postures will be developed to show the cost and effectiveness implications of decisions on these issues Because it is assumed that the U S would not want to depend on preemption to accomplish our fundamental strategic force planning objectives the effectiveness of alternatiVe forces will be evaluated in tents of their survivable retaliatory capability Naturally when examining alternative force postures from the viewpoint of perceptions and crisis stability first and second-strike capability will be addressed The strategic options matrix Figure l from the study outline was developed to address each of these issues The objective of this paper is to seek Secretary of Defense approval for 1 our proposed changes to the matrix which reflect our better understanding of the issues and scope of the study and 2 our proposed approach for conducting the cost effectiveness analysis smears Recall that each column and row in the original matrix represented strategy decisions the Administration could make regarding the amount of non-silo retaliatory capability appropriate for the ICBM force represented by the columns and the characteristics of the desired hard-target kill capability represented by the rows He are proposing that the strategy decisions represented by the columns regarding the amount of non-silo capability of the ICBM force be changed from the Relative Role of to Do He Want An Effective Triad do we want to modernize the ICBM force so that all three legs of the Triad make a significant contribution to the retaliatory capability of our strategic forces see Figure 2 This statement better reflects the immediate issue for decision which is 59 the ultimate deployment level of MX but whether or not we should modernize the ICBM force with a more survivable missile in order to retain an effective Triad The refinemeht of the ultimate size of the ICBM force can be left to be worked out after SALT is settled and the review of our targeting objectives has been completed SUBJECT 2 0 EXECUTIVE ORDER Figure - mm mi BASIC Relative Role ICBMs ICBHs Play a of ICBMs _Hinor Role 1 - 2 3 4 Balanced Triad a with Balanced Triad with uwivable More ICBM Capability ICBH Capability Unbalanced Triad ard Target -than Current Comparable to Transitional with and Capability orces lToday's Forces a Region Bombers Dominant Extensiveb Quick- Response Capability Extensive Non-Quick Response Capability No First Strike Capa- bility ainst Silos Lin ted Countersilo Cppahility Balanced Triad Each leg has the survivable capability to destroy a significant portion of bthe specified non-silo target base see Section Extansivo'caapability Capability to destroy a largo percentage of silos and other hard targets nrnan OLURLI FigUn 2 Revised Strategic Options Matrix DoMe Want An Triad with udder-nixed Effective Triad ICBM Force 'Survivabie Hard Target Unbaianced Triad with No- ICBM Modernization Extensive Quick Response Capability ExtensiverNon Quick- Response CapabiTity NO First-Strike Against SiTos Limited Counter- Siio Capability 4 Within each square in the matrix there will be several alternative force structures that satisfy the conditions represented by the respective column and row headings For example three ways to satisfy the conditions of square lA are l A Triad where the ICBM force has roughly the non-silo retaliatory - ca ability of today assuming the planning factor for plus sufficient capability to do all of the countersilo mission no 13526 section 2 A Triad where the non-silo retaliatory capability of the ICBM force is roughly the same as that of the future SLBM or bomber forces plus sufficient capability to do all of the countersilo mission 3 Same as 2 above only the countersilo mission is performed by the SLBM force within Column 1 we will examine the relative merits of effective Triads where the emphasis is placed on ICBMs or bombers as well as combinations of these force elements For Column 2 the assumption is that the ICBM force will not be modernized with a more survivable system C CostuEffectiveness Analysis After careful consideration we have decided that it would be impractical to conduct the cost-effectiveness analysis using either equal effectiveness or equal cost forces Here's why 9 With respect to equal effectiveness forces Without a definitive statement of our targeting objectives we have develOped a methodology see Appendix B that allows us to depict the capability of alternative forces across a wide range of economic and non silo military target destruction levels This measure of effectiveness is not a point but rather a curve - - Since the force alternatives MX Trident II ALCMs would be deployed over different time frames and at different rates the effectiveness of any one force posture will vary with-time Hence the effectiveness curve is really a surface - - In addition we plan to evaluate the effectiveness of alternative force postures in terms of their hedging potential as has been discussed in another section of this interim report Clearly equal effectiveness forces at this level of detail is an analytical impossibility -0 With respect to equal cost forces - Disagreements abound over whether constant or discounted cost analysis is most appropriate ll l k' ll OLUHLI 5 -- Discounted dollar analysis loses some of its utility when capability is delivered in substantially different time profiles f Constant dollar comparisons are just discounted costs using a zero discount rate There are disagreements over whether 20 30 or more years is the appropriate costing horizon At the very least a number of different discount rates and time periods would be required in order to fend off criticism that the study was biased one way or the other This would require generating entirely new force structures for each discount rate or costing period Not only would this quadruple or more the work but more importantly it would make the effectiveness analysis and in particular the structuring of force mix alternatives dependent on costing assumptions Finally and most importantly equal cost or equal effectiveness forces do not represent the alternative decision choices we actually face A more realistic set would be the various ways we could structure our forces to meet the SALT THO limits Accordingly we have decided on the following approach which we believe is more apprOpriate for this study Using the new Options matrix Figure 2 a number of forces will be develOped for each square in the matrix These forces will be programmatically feasible and consistent with the numerical limits of the SALT THO agreement extended into the 19905 The effectiveness of these forces will be compared using the methodology described in Appendix B Each of the forces will be casted for a number of discount rates including zero through the year 2000 In some cases the relative cost-effectiveness of alternative forces may depend on the measures of effectiveness and or the costing approach considered By presenting a range of effectiveness results and a variety of discount factors decision-makers will be able to see the sensitivity of the results Hopefully this will facilitate the decision making process In order to keep the number of forces analyzed to a manageable number will develop a set of representative forces for each square in the matrix These forces will be presented to the study working group for review and coment To the extent possible we will arrive at an agreed set of forces for the cost-effectiveness analysis We will limit the number of forces for each square to the minimum necessary to highlight the range of available options Conclusions will be presented about the relative cost effectiveness of forces within a given square as well as among different squares The latter will be used to highlight the costs and effectiveness of will take into account nuclear material availability realistic deployment rates etc different approaches to satisfying 1 a given strategic option same row or column and different strategic options different row and column Excursions be important and invoive 1 advanced threats and unanticipated defensive breakthroughs 2 seiected SALT possibilities greater reductions breakdown and 3 severe1y constrained budgets These excursions wi11 be conducted on a 1imited number of forces either those that appear to be the most cost-effective or those that appear to be the most sensitive to the various excursions Once again to the extent possible agreement Hi1 be sought in the working group on which forces and which excursions are to be conducted where agreement cannot be reached a1ternate points of View be forwarded with the final report to the Executive Study Committee APPENDIX B ANALYTICAL METHODOLOGY 1 Introduction This paper describes the analytical methodology proposes to use in the MX Triad study to evaluate the military capability of alter- native U S arsenals for the following tasks targeting the Soviet Union targeting the People's Republic of China PRC and the non-Soviet Harsaw Pact NSHP and maintaining a Secure Reserve Force SRF Only the targeting of the Soviet Union is dealt with in detail In describing capabilities against the Soviet Union distinctions are drawn among economic targets all other non-silo targets in the data base primarily military installations - ICBM silos We believe it is important to describe an arsenal's capabilities against each of these target groups However dealing with the large group of non-silo non-economic targets is the most difficult technically and this problem is given the most analytical attention in the present paper has made a considerable effort in recent months to elucidate long-standing differences in weapon requirements determined by simple allocation models and SIOP procedures As a result of analysis and discussions with JCS and the Air Force we have developed three constraints for use with the model These constraints are not intended to replicate their SIOP anaiogues but rather to introduce in an approximate way considerations that clearly play a major role in operational targeting The constraints are - a requirement that weapons be allocated in waves a first wave against military targets and if appropriate a second wave against economic targets - a requirement that the more important non silo non-economic targets be attacked to a specified aggregate before targets of lesser priority are directly attac an E_0_1352635ec 0n I a requirement in some calculations that certain classes of targets be redundantly covered so as to assure a hedge against failure of a portion of the Triad All of these constraints have the effect of increasing weapon requirements because ail of them create special objectives to be met in addition to high levels of aggregate DE In the absence of the constraints the model using current SIOP forces against the current data base h - predicts that the aggregate damage levels achieved by the SIOP can be achieved with 30% fewer weapons than used in the the above constraints the differences in weapon count are in the noise of the calculations e g differences of Although the amended alloca- tion rules are by no means a simulation of the SIOP there are major differences in weapon allocations at the microscopic level we believe it now includes the features necessary for it to be useful in the MX Triad study we can now examine issues such as the military implica- tions of increased silo vulnerability more clearly primarily because we now can focus on several classes of military targets and explore the rissue of hedging more informatively The paper previews some issues related to the methodology and suggests several areas where the proposed methodology can be augmented by analytical support from SAGA AF SAC a SA and the Navy We hope that many of the concerns some had with the analytical approach used by in the Consolidated Guidance have been eliminated but we must continue to recognize that models appropriate for force-mix' studies must necessarily be simplified representations of reality The following describes our approach to dealing with each of the targeting requirements beginning with the most straight forward PRC NSNP and SRF and concluding with the more complex Soviet non silo II PRC NSWP and SRF Requirements One of the major criticisms of the calculations in the Consolidated Guidance CG was that no forces were allocated to cover targeting requirements for the PRC the NSNP and the SRF For each force posture evaluated in this study forces will be set aside to cover these require ments These forces will be comparable to today 5 level of capability under alert with damage and generated without damage scenarios Today the following forces are identified to satisfy PRC NSHP and SRF targeting objectives Number of Weapons Number of weapons Day- to Day Alert Generated Alert System NSHP PRC SRF NSHP FEE SRF B- Gravity Bomb 11 SRAM Minuteman Poseidon Polaris 131 0 13526 section 3 3001 5 1 MOD Alert BM warheads 2 Includes HOD Alert SLBM Harheads 3 We do not intend to analyze whether or not these forces are sufficient but rather we uill set aside from each force posture comparable forces to satisfy these requirements Countersilo Requirements In order to fully address the countersilo issue lev ls of counter silo capability well in excess of that achievable with current forces will be considered This issue will be treated for the most part independently from the Triad question although clearly a decision on the amount of countersilo capability desired will have a major impact on our future force structure Countersilo capability of alternative systems MX Trident II cruise missiles will be compared in a straight forward manner Calculations will be made of the number of weapons required and the cost to satisfy a range of damage expectancy DE goals under various assumptions regarding SALT and alert status Inputs for these calculations will be provided by U S weapon system characteristics and DIA projected Soviet ICBM target system characteristics Sensitivity analyses will be conducted fer the key parameters The output of this effort will be weapons requirements costs and uncertainties It appears unfortunately that a couple of key questions important to the countersilo issue will remain qualitative and hence highly judgemental Two examples come immediately to mind the time required for the RCA to get the Emergency Action Message EAM out may be significantly greater than the differences in reaction time of the various legs of the Triad and 2 the time urgency of Soviet silos as a function of scenario These and other considerations bear directly on the countersilo issue They will be discussed quantitatively to the degree possible and all sides of the issue will be outlined at least qualitatively in parallel with the calculations of weapon requirements IV Non-Silo Requirements The approach proposed to evaluate the non-silo capability of alter- native U S forces uses as a baseline the level of capability in our Current forces as they are actually employed in current targeting plans SIOP-SB This is not to say that our current capability is or should be a recognized standard of sufficiency nor that the current arsenal and the way it is employed provide a prOper standard for measuring future force requirements There is in fact no basis at this time for predicting what our future targeting or secure reserve force policy might be or what forces would be required for its implementation Future targeting and SRF objectives are being considered independently in directed studies which are proceeding in parallel with this study At this time we probably cannot expect these studies to define future base case targeting objectives or SRF requirements for use in our analysis He propose to use our current capability to satisfy non silo objectives as a baseline for relative comparisons alternative force postures will be nonmalized to the current capability of SIOP-SB forces because of the need to consider the capability of alternative forces on the basis of some known standard 4 The prOposed methodology will be described below using as an example the calculation and display of current force capability The discussion will consider four major topics - Arsenal description Target base - Allocation rules - Display of results l Arsenal Description During the course of the study alternative force mixes will be constructed to satisfy the combinations of hard target kill and level of ICBM capability contained in the options matrix illustrated below In each case the number of systems available for targeting non-silo installations in the USSR will be determined by 1 ap lying appropriate alert rates to on-line or unit effective forces 25 setting aside forces to satisfy a countersilo requirement if appropriate and 3 setting aside forces to satisfy PRC NSHP and SRF requirements as discussed in Section II Available systems will be described by weapon system characteristics yield CEP and weapons system reliability and by pre-launch PLS and penetration probability FTP factors calculated in separate analyses reflecting DIA threat data It appears likely that sensitivities to changes in probability of arrival PA WSR PLS PTP will have to be- explored parametrically since future threat descriptions are necessarily highly uncertain - Strategic Options Matrix no He Hunt An Triad Hith Hodcrnized Unbalanced Triad Hith No Effective Tried ICBH Force ICBH Modernization Survivable Herd Target Capability Extensive Quick- Response Capability Extensive Hon Quick- Response capability No First-Strike Capability Against Silos Limited Counter- Silo Capability 5 The foiiowing tabie describes the forces avaiTabie in today's SIOP-SB arsenai to target oniy non-silo instaiiations in the Soviet Union under the alert with damage AND and generated without damage GHOD scenarios These are the forces that wi11 be used to calculate the baseiine capabiTity against which aTternative future forces wiTT be compared Portions denied are S-FRD and 'thus outside of the jurisdiction of the Intel-agency Security Classi cation Appeals Panel AWE AND GWUD GNOD Name egg Yier 1 Warheads PA g Harheads At Sea Poseidon In Port Poseidon ATert Gravity Bombs AT ert SRAMs NonwAiert Gravity Bombs Non Aiert SRAM Titan 1 The arsenai characterizes bomber weapons in terms of the two most predominant yieic for purposes of simpiicity 2 PA PLS NSR FTP 3 Refiects average number of off-Tine at sea and therefore survivabie 13526 section 3 303 5 7 2 Target Base Description The Soviet target base refTects the February 1977 NTB The detaiied NTB instaiiation data is processed for by the Command and Controi Technicai Center CCTC as foiTows - The first step in the process is to break the instaiiation data base into independent groups or compiexes of instailations The purpose of this step isfto group instaTiations into mane eabTe subsets m his Tito 13526 section - The next step in the process is to attack the instaTlations in each com Tex This attack is carried out untiT a specified Portions denied are S-FRD and thus outside of the jurisdiction of the Interagency Security Ciassi cation Appeals Panel I 13 0 13526 section damage level is achieved on each installation recess la 5 down the ins a a ions amage from this weapon Those exceeding the level are assigned to this 062 and are excluded from affecting the creation of subsequent 0625 Subsequent weapons are similarl applied Once all the installations are damaged to at least the level aim areas are generated by assigning each installation to the 062 which actually damaged it the mozti These aim areas are the basic targets used in the allocation mo Once aim areas are defined Yields are chosen that bracket the range i i and may be different than either the complexing or aim area generat on yields For response information generation each aim area is assumed to be independent no collateral damage between aim areas weapons are allocated to each aim area to achieve a DE of at least for all included installations The actual damage achieved an ave ge weapon radius a function of the yield and individual installation vu erability and the number of weapons required to achieve the damage a determined and saved me 13526 section - Finally the value contained in each aim area by attack objective category recovery military leadership etc is filed along with the attack results listed above with this information enough is known about each aim area to be able to use them as targets in the allocation model When using aim areas as targets for the allocation models makes a number of simplifying assumptions regarding the way in which installations within the aim area will respond to nuclear detonations The first assumption is that when an aim area is attacked by any weapon all installations within the aim area are attacked and respond in an identical manner A second assumption is that the aim areas respond as point targets This assumption is necessary in order to easily predict the damage response of the aim area to the detonation of a real weapon reliability is less than 1 0 and CEP is greater than 0 The iresponse functions can be treated as simple one parameter fits The response function has the form '1 7 where PD probability of damage applied identically to all value classes within the aim area number of weapons fit parameter based on the response information A third assumption is that the parameter can be found by interpolation from the yields and destruction probabilities found in the description of the aim area for any yield in the range defined by the four nominal yields While mathematics is used extensively in the aim area creation process calibration runs must be performed to determine that the aim areas and their response to real weapons adequately represent the original NTB installation data base These comparisons with CCTC results indicate that weapon re uirements against the aggregated target base are more conservative requires more weapons than generated requirements from the detailed source data especially at low levels of damage This conservatism in the approach is due to the aggregation that loses visibility to a small number of very high payoff targets The modeled target data base currently has fidelity to SIDP-SB objective categories of targets leadership recovery etc that is levels of damage can be Specified for each objective category and damage including collateral damage to non-objective installations can be assessed for each category of targets As in the SIOP economic recovery damage objectives are specified and results are displayed in terms of economic recovery value average of manufacturing value added and capital cost to replace For non-economic installations damage objectives and reported results are in terms of the percent of installations destroyed Since we have no value scheme to guide the destruction of non-economic installations as in the economic recovery case the data base flags a subset of non-economic installations which have been identified as being militarily more important The subset excludes the silos but includes hardened launch control facilities LCF LCFs could be considered under the ICBM targeting requirement but have been retained here to stress the hard target capabilities of available systems and to provide a hedge against future hardening of other installations in the data base 19 0 13526 section As will be deScribed below the allocation rules or this ana ysis have been established such that the specified levels of aggregate DE by objective are achieved preferentially against the militarily more important subset in the is a sequentially optimal DGZ generator that has full fidelity to primary and collateral damage against all elements of the Target Bate Base NTB process of attacking the entire non-silo NTB target base In addition a DE level is specified against more important hard nuclear threat installations primarily launch control centers to insure that these installations are not avoided in achieving aggregate DE levels The number of non-economic installations in'the data base is summarized in the following table In the data base each installation appears only against the objective where it has the highest priority The single counting of NTB installations accounts for the significant differences in and installation counts The counts the same installation and its destruction in as many objectives as it may appear when counted in the same manner and counts are similar Number of Non-Economic Installations 1 Data Base Data Base More All Adjusted for Objective Important Less ICBM Multiple Objectives Resource Recovery Military Leadership Nuclear Threat less ICBM Conventional Force TOTAL E O 13526 section 1 Economic recovery installations and their associated recovery value are of course also included in the data base 2 The difference in Objective military installations is attributed to changes in the data base since February l977 3 weapon Allocation Rules Over the past several months we have made an effort to gain an understanding of how the current SIOP is constructed and how it would be carried out We have incorporated features in our model that permit us to more faithfully duplicate the way in which weapons are actually employed in the SIOP and in the process our results have closely approximated consequences of execution Current employment guidance_ identifies Major and Selected Attack Options and SAOs and specifies that these options should be encom assed in one inte rated plan im act of this IE O 13526 section I 112 0 13526 section IQ SEBHEI 9 A second allocation constraint is that we are requiring specified levels of damage expectancy by objective against a militarily more important subset of non economic installations This is consistent with prudent military planning but is generally not done in allocations designed to minimize warheads to achieve high overall aggregate damage expectancies A maximum DE level of chosen to provide a reasonable fit of the calculated results with the current SIOP consequences of execution no 13526 section A third constraint is that we are requiring that an equal percentage of each available system in the arsenal be used to achieve the specified goals against the non-economic recovery installations we do this_to prevent the model from choosing all of the better weapons in the arsenal in the allocation leaving only less capable systems for the MAO-2 increment This is somewhat comparable to the military planner setting aside a balance of weapon types for the economic recovery task and Optimally allocating his remaining arsenal against non-economic installations The allocation rules and the attack sequence used to calculate the current force baseline are sunmarized as follows These same rules will also apply to the calculation of the relative capability of alternative future forces E0 13526 section I WET- E 0 13526 section 3 3031 5 With the approach described above it is also possible to calculate the inherent capability of individual elements of the total force or combinations of those elements to achieve damage across a spectrum of economic value versus non-economic non-silo installation damage expectancies For example we can calculate using the same procedures outlined above the capability of the ICBM force alone or the ICBM force in combination with bombers or This approach assumes an optimum allocation of the weapon systems specified -e that is the planner would have to know that the other legs were unavailable I 12 0 13526 section 11 and retarget the available arsenal to achieve the calculated results believes that this approach offers promise as a means of showing the maximum hedge potential of a given total force structure and as a means of illustrating the flexibility of systems or legs in the force mix to achieve damage levels across the non-silo target Spectrum 4 Display of Results This section discusses the display of force capability to satisfy non-silo Soviet targeting objectives In all cases the results assume that weapons required for other objectives have been set aside At this point in our study effort we plan to display the results of comparative force analyses across the spectrum of possible Soviet targeting objectives This has the advantage of showing the sensitivity of the force mix to alternative targeting policies It also shows the capability of the force mix across the full non-silo target spectrum Figure 1 below shows the calculated capability of the SIOP-SB AND non-silo USSR only arsenal given the detailed arsenal target base and allocation rules described above The figure shows 1 the tradeoff between economic recovery value and non silo non-economic installation destruction for example when the objective is 70% DE against non-silo military and political installations along the horizontal axis the corresponding DE achievable against economic recovery value is about 68% and 2 the high cost associated with damage expectancies greater than 80% for economic recovery targets and 60% for the total set of non-silo military and political installations The on the chart approximates the Consequences of Executing SIOP-SB Figure SIDP-SB AHD Forces 80 SIOP-SE Consequences of Execution ao_ 40 1 Economic Recovery Voice Husband 2ohut-Iii $1130 and hiitml Installations Incl-eyed E0 13526 section r Using the same basic display we can show the calculated capabilities of combinations of Triad force elements and of the individual legs as in Figure 2 below in each case these results represent an optimum reallocation of the available element s and therefore correspond to the maximum degree of hedge potential and flexibility available in the force element s The figure shows 1 the dominance of the ICBM the assumed in SIOP-SB the higher yield and better accuracy associated w1 -o ay's and 2 the lack of capability in the current AHD USSR nonwsilo only SLBM and bomber forces individually and in combination to achieve high damage expectancies greater than 50% against non-silo non-economic targets force across the siectrum of possible targeting objectives reflecting Figure 2 SIOP-SB AND Forces Non $310 USSR Only I00 'c s Mir-tic so W51 40_ BD sl 1 Economic Recovery Value Destroyed ml 20 lb 40 51' SG 100 I of howSilo Hilton-y and Political Institutions Instr-om It should be noted that the model does not have good fidelity I 13 0 13526 section II IV 13 We plan to present the capability of alternative future forces in the same way as shown in Figures 1 and 2 with current force capability as a reference It is possible that some future force mixes might be so capable as to exceed the boundaries effectively at 90% in both dimensions of the damage expectancy matrix and still have forces left over To account for this possibility and also to provide an alternative means of displaying the capability of future forces relative to current capability we propose to select points along the DE curve for our current forces Figure 1 that are representative of damage objectives across the spectrum -- 90% economic value 45% other 76% economic value 68% other 65% economic value 72% other and 45% economic value and 80% other 2 calculate the minimum fraction that when applied uniformly to each system in the future force mix will provide forces with exactly enough capability using the methodology described above except that the economic recovery objective is fixed to achieve the specified set of goals and 3 plot the inverse of the calculated fraction or force multiplier that is Relative Force Size for purposes of relative force comparisons The following'table summarizes an example of this approach using illustrative Relative Force Size calculations inverse force multipliers that could-represent the capability of some future forces Figure 3 illustrates how the data would be displayed DE Objectives Relative Force Size Economic gthg - fg g_g forgg_ 90% 45% l 3l l 23 l 23 76% 68% 1 35 - 1 18 l l2 ass 72% 1 35 1 15 1 07 45% 80% 1 41 1 11 97 IV methodology Issues A number of potential issues are not addressed specifically by the basic methodology described in the previous section While the methodology is well suited to define relative force capabilities and attributes in large scale exchanges against a large data base certain details sensitivities and fact-of-life operational constraints cannot be readily accounted for in a highly aggregated fast running model Many of these items can be treated explicitly but external to the basic model others are judgmental and more philosophical in nature and difficult to treat explicitly The non-silo USSR portion of the analysis is important to the study and the results of this portion should be most closely scrutinized However the models involved are sufficiently complex that the details of the methodology will not be well understood by the casual reader This makes it vital that the limitations and uncertainties be clearly discussed The details of the model and its level of aggregation has anemones Figure 3 ReTative Force CapabiTity to Destroy Non-$510 Soviet Targets 2 0 More Eco Recovery Current Emphasis - Emphasis 2-0 Eco 68% Other 0 90% Eco 45% OtherSIOP ALT ALT ALT SIOP ALT ALT Less Eco Recovery Non-economic Emphasis 2-0 Emphasis 2 0 45% Eco 803 Other 65% Eco 72% OtherSIOP ALT ALT ALT SIOP ALT ALT ALT each case warheads to satisfy requirements have been set aside and are not refTected in the relative capabilitv shown I ED 13526 section I is been discussed above However there are several areas which required further discussion and analysis - First the study is examining force structure for the mid-805 and beyond It is unclear how the target system will change Certainly there will be economic growth with an attendant increase in the number and type of economic installations as well as a reranking of the value structure within existing installations This growth may or may not result in a requirement for more DGZs to destroy a comparable percentage of recovery value Similarly a concerted effort in Civil Defense could have a significant impact on both the size of the target system dispersal and the hardness of the_installations There might be a trend toward more hard targets even without a concerted Civil Defense effort A change in targeting guidance could have a similar effect -- for example a change that emphasized high levels of destruction of command control and communications installations _ Because of the lack of hard data in these areas we are currently uncertain as to how much we can do quantita tively in this study At the least we can discuss possible impacts qualitatively - Second the analysis will uSe the February 1977 as the installation list for developing aim areas Changes in the NTB since that time include 1 eliminating the economic recovery circles 5-955 and replacing them with economic recovery installations and 2 reordering the primary objective recovery leadership etc associated with some installations The effect of these changes has not been assessed in detail but should have minor impact on results based on the experience with similar changes in the past Third the aggregated model does not account for the detailed operational constraints inherent in an actual SIOP laydown These constraints result in a less than optimal allocation and include -- Range -- all targets cannot be reached by all weapons -- Fodtprint Soviet target location does not necessarily match the footprint capability of our strategic systems Even bombers may be affected This may be further aggravated for a multiple weapon system by polic constraints such as option booster and country purity optiona withholds and minimization of collateral damage -- Time urgency considerations may require the assigning of specific weapons to specific targets It has not been possible to quantify the significance of the degrade in targeting efficiency due to these constraints Even if a number could be assigned which reflected the current SIOP allocation it would not be representative of future forces with different system characteristics However in the application of this model several is currently processing the May l978 NTB for use in PAEE analysis However this data not be available in time to be used in this Study inner-nap factors result in an overstatement of the weapons required and these tend to compensate for not treating the operational constraints explicitly In addition the model will not be used for absolute force sizing but rather for a relative comparison of alternative forces the ranking of which will not be measurably affected by these small differences Fourth military targets are all of equal value in the Soviet targeting analy51s as in most analyses although a higher priority set of non-economic targets is defined However in the SIOP world where the allocations are hand massa ed a sub ective non-unantif' - scheme is apparent n 'n approach described in He cannot a and w 0 not a ieve at th evel of detail is appropriate or necessary for study 13526 section V Analytical Support From Other Agencies The preceding sections have suggested several areas where separate analyses will be required in support of the overall study objectives This section will briefly summarize areas where additional analyses will probably be required and agencies presently capable of undertaking them A Planning Factor Analyses PLS and PA calculations will need to be derived using BIA-provided'threat data Alternative threat levels will be treated parametrically These calculations may require Air Force SA and SAC and Navy 604 support B Operational Constraints As was described above range footprint and option purity are not explicitly addressed in the method- ology preposed for this study It may be desirable to test the sensitivity of some alternative forces mixes to these constraints Air Force SAC SAGA and Navy UP 604 have capabilities to address some or all of these considerations C Measures of Merit The methodology proposed in this paper has addressed a measure of capability to destroy a given set of targets It has not addressed cost effectiveness considerations nuclear material considerations or minimum cost solutions where cost could be in terms of MIRVs etc J These are issues which will bear strongly on the construction of specific force structure alternatives Considerable capability to address these issues rests in DOD agencies Navy in particular has a model that can derive minimum cost solutions to achieve alternative levels of damage against both economic and_other military installations This model was used extensively in analysis WEE-F 7 D ed in The maximum hedging potentiai of aiternative force mixes was i11ustrated in Section IV as a of the proposed methodoiogy However because the hedging issue promises to be criticai to this -study more detaiied analyses of this issue wiil be required has developed a methodoiogy -- actuaiiy a set of force a11ocation rules -- which drives the mode1 to achieve pre-pianned hedging of economic and or other instaliations In addition SAGA AF and Navy aiso have analyticai to address the hedging issue This document is from the holdings of The National Security Archive Suite 701 Gelman Library The George Washington University 2130 H Street NW Washington D C 20037 Phone 202 994-7000 Fax 202 994-7005 nsarchiv@gwu edu