The Analytic Network Process. Examples. Part 2.3

We give here the details of an application of the ANP to National Missile Defense (NMD), a $60 billion decision (one of the costliest undertaking ever) by the United States to deploy an anti-nuclear-missile missile defense system. This study I did in September 2000 developed with its strategic crite...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Datum:	2003
1. Verfasser:	Saaty, T.L.
Format:	Artikel
Sprache:	English
Veröffentlicht:	Навчально-науковий комплекс "Інститут прикладного системного аналізу" НТУУ "КПІ" МОН та НАН України 2003
Schriftenreihe:	Системні дослідження та інформаційні технології
Schlagworte:	Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах
Online Zugang:	https://nasplib.isofts.kiev.ua/handle/123456789/50299
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!
Назва журналу:	Digital Library of Periodicals of National Academy of Sciences of Ukraine
Zitieren:	The Analytic Network Process. Examples. Part 2.3 / T.L. Saaty // Систем. дослідж. та інформ. технології. — 2003. — № 4. — С. 7-23. — Бібліогр.: 3 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine

id	nasplib_isofts_kiev_ua-123456789-50299
record_format	dspace
spelling	nasplib_isofts_kiev_ua-123456789-502992025-02-23T18:01:19Z The Analytic Network Process. Examples. Part 2.3 Аналітичні мережеві процеси. Приклади. Частина 2.3 Аналитические сетевые процессы. Примеры. Часть 2.3 Saaty, T.L. Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах We give here the details of an application of the ANP to National Missile Defense (NMD), a $60 billion decision (one of the costliest undertaking ever) by the United States to deploy an anti-nuclear-missile missile defense system. This study I did in September 2000 developed with its strategic criteria, and benefits, opportunities, costs and risks. Deploy NMD is the highest priority outcome, that is validated by sensitivity analysis. It was presented to the National Defense University in Washington in February 2002. In mid-December 2002 the US Government announced that it planned to develop the NMD — at best an interesting coincidence. Another example, done in September-October 2002, was about the strategy that the US should follow to remove the Saddam Hussein regime in Iraq. Going with the UN was the best outcome followed by a going alone or with a coalition. Викладаються подробиці застосування АМП до Національної протиракетної програми — рішенню США створити ракетні протиядерні системи (NMD). Програма оцінювалась у $60 млн (одне з найдорожчих реалізацій в історії). Це дослідження, яке проведено автором у вересні 2000 р., базувалося на стратегічних критеріях з урахуванням ефективності, можливостей, вартості і ризику. Розгортання NMD — найвищий пріоритетний результат, що підтверджено виконаним аналізом. Робота була подана у Національний університет оборони (Вашингтон) в лютому 2002 р. В середині вересня уряд США зробив заяву, що він планує створення NMD. Другий приклад, розглянутий в вересні–жовтні, торкається стратегії США в усуненні режиму Саддама Хусейна в Іраку. Одержано висновок, що найкращим варіантом рішення є сумісні дії США та ООН. У крайньому же випадку допускалися дії поодинці або у коаліції з іншими державами. Излагаются подробности применения AСП к Национальной противоракетной программе — решению США создать ракетную противоядерную систему (NMD). Программа оценивалась в $60 млн (одно из самых дорогостоящих реализаций в истории). Это исследование, выполненное автором в сентябре 2000 г., базировалось на стратегических критериях с учетом эффективности, возможностей, стоимости и риска. Развертывание NMD — наивысший приоритетный итог, что подтверждено выполненным анализом. Работа была представлена в Национальный университет обороны (Вашингтон) в феврале 2002 г. В середине сентября 2002 г. правительство США объявило, что оно планирует создать NMD. Другой пример, рассмотренный в сентябре–октябре 2002 г., касался стратегии, которой должны придерживаться США для устранения режима Саддама Хусейна в Ираке. Получен вывод, что наилучшим вариантом решения являются совместные действия США и ООН. В крайнем же случае допускались действия в одиночку или в коалиции с другими государствами. 2003 Article The Analytic Network Process. Examples. Part 2.3 / T.L. Saaty // Систем. дослідж. та інформ. технології. — 2003. — № 4. — С. 7-23. — Бібліогр.: 3 назв. — англ. 1681–6048 https://nasplib.isofts.kiev.ua/handle/123456789/50299 519.5 en Системні дослідження та інформаційні технології application/pdf Навчально-науковий комплекс "Інститут прикладного системного аналізу" НТУУ "КПІ" МОН та НАН України
institution	Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection	DSpace DC
language	English
topic	Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах
spellingShingle	Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах Saaty, T.L. The Analytic Network Process. Examples. Part 2.3 Системні дослідження та інформаційні технології
description	We give here the details of an application of the ANP to National Missile Defense (NMD), a $60 billion decision (one of the costliest undertaking ever) by the United States to deploy an anti-nuclear-missile missile defense system. This study I did in September 2000 developed with its strategic criteria, and benefits, opportunities, costs and risks. Deploy NMD is the highest priority outcome, that is validated by sensitivity analysis. It was presented to the National Defense University in Washington in February 2002. In mid-December 2002 the US Government announced that it planned to develop the NMD — at best an interesting coincidence. Another example, done in September-October 2002, was about the strategy that the US should follow to remove the Saddam Hussein regime in Iraq. Going with the UN was the best outcome followed by a going alone or with a coalition.
format	Article
author	Saaty, T.L.
author_facet	Saaty, T.L.
author_sort	Saaty, T.L.
title	The Analytic Network Process. Examples. Part 2.3
title_short	The Analytic Network Process. Examples. Part 2.3
title_full	The Analytic Network Process. Examples. Part 2.3
title_fullStr	The Analytic Network Process. Examples. Part 2.3
title_full_unstemmed	The Analytic Network Process. Examples. Part 2.3
title_sort	analytic network process. examples. part 2.3
publisher	Навчально-науковий комплекс "Інститут прикладного системного аналізу" НТУУ "КПІ" МОН та НАН України
publishDate	2003
topic_facet	Проблеми прийняття рішень і управління в економічних, технічних, екологічних і соціальних системах
url	https://nasplib.isofts.kiev.ua/handle/123456789/50299
citation_txt	The Analytic Network Process. Examples. Part 2.3 / T.L. Saaty // Систем. дослідж. та інформ. технології. — 2003. — № 4. — С. 7-23. — Бібліогр.: 3 назв. — англ.
series	Системні дослідження та інформаційні технології
work_keys_str_mv	AT saatytl theanalyticnetworkprocessexamplespart23 AT saatytl analítičnímereževíprocesiprikladičastina23 AT saatytl analitičeskiesetevyeprocessyprimeryčastʹ23 AT saatytl analyticnetworkprocessexamplespart23
first_indexed	2025-11-24T07:39:20Z
last_indexed	2025-11-24T07:39:20Z
_version_	1849656565201633280
fulltext	© Thomas L. Saaty, 2003 Системні дослідження та інформаційні технології, 2003, № 4 7 TIДC ПРОБЛЕМИ ПРИЙНЯТТЯ РІШЕНЬ І УПРАВЛІННЯ В ЕКОНОМІЧНИХ, ТЕХНІЧНИХ, ЕКОЛОГІЧНИХ І СОЦІАЛЬНИХ СИСТЕМАХ UDC 519.5 THE ANALYTIC NETWORK PROCESS. EXAMPLES. PART 2.3 THOMAS L. SAATY We give here the details of an application of the ANP to National Missile Defense (NMD), a $60 billion decision (one of the costliest undertaking ever) by the United States to deploy an anti-nuclear-missile missile defense system. This study I did in September 2000 developed with its strategic criteria, and benefits, opportunities, costs and risks. Deploy NMD is the highest priority outcome, that is validated by sensitivity analysis. It was presented to the National Defense University in Washing- ton in February 2002. In mid-December 2002 the US Government announced that it planned to develop the NMD — at best an interesting coincidence. Another exam- ple, done in September-October 2002, was about the strategy that the US should fol- low to remove the Saddam Hussein regime in Iraq. Going with the UN was the best outcome followed by a going alone or with a coalition. 1. INTRODUCTION — THE NATIONAL MISSILE DEFENSE (NMD) Having given the steps of the Analytic Network Process (ANP) in Part 2.2, we will dedicate Part 2.3 to full blown examples of applications of the ANP. Not long ago, the United States government faced the crucial decision of whether or not to commit itself to the deployment of a National Missile Defense (NMD) system. Many experts in politics, the military, and academia had ex- pressed different views regarding this decision. The most important rationale be- hind supporters of the NMD system was protecting the U.S. from potential threats said to come from countries such as North Korea, Iran and Iraq. According to the Central Intelligence Agency, North Korea’s Taepo Dong long-range missile tests were successful, and it has been developing a second generation capable of reach- ing the U.S. Iran also tested its medium-range missile Shahab-3 in July 2000. Op- ponents expressed doubts about the technical feasibility, high costs (estimated at $60 billion), political damage, possible arms race, and the exacerbation of foreign relations. The idea for the deployment of a ballistic missile defense system has been around since the late 1960s but the current plan for NMD originated with Presi- dent Reagan’s Strategic Defense Initiative (SDI) in the 1980s. SDI investigated technologies for destroying incoming missiles. The controversies surrounding the project were intensified with the National Missile Defense Act of 1996, intro- duced by Senator Sam Nunn (D-GA) in June 25, 1996. The bill required Congress to make a decision on whether the U.S. should deploy the NMD system by 2000. The bill also targeted the end of 2003 as the time for the U.S. to be capable of de- ploying NMD. The idea explored in this project is to develop and illustrate the three phases with a timely example, the intricate and very costly decision regard- ing a National Missile Defense (NMD) system. Because of the possibility of de- Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 8 pendence and feedback, we use the Analytic Network Process (ANP) and its software Super-Decisions with its sensitivity analysis option to examine the NMD decision. On February 21, 2002 this author gave a half-day presentation on the subject to National Defense University in Washington. In December 2002, Presi- dent George W. Bush and his advisors decided to build the NMD. This study may have had no influence on the decision but still two years earlier (September 2000) it had arrived at the same outcome. The alternatives we considered for this analy- sis are: Deploy NMD, Global defense, R&D, Termination of the NMD program. 2. CRITERIA AND DECISION NETWORKS The second column of Tabl. 1 shows the criteria of each BOCR. For example, there are four benefits criteria: Economic (0.157), Political (0.074), Security (0.481) and Technology (0.288). The priorities attached to each are obtained through pairwise comparisons. Each criterion under benefits has subcriteria such as Local Economy and Defense Industry under Economic. Again, the priorities of the two subcriteria are obtained from pairwise comparisons and similarly for the remaining criteria and subcriteria under opportunities, costs and risks. Opportuni- ties and risks have no subcriteria. The total number of criteria and subcriteria used as control criteria for the comparisons made in the networks is 23. The global pri- orities of these criteria (subcriteria) shown in the last column of Tabl. 1 are ob- tained by weighting their priorities by those of their parent criterion if there is one. For example, for local economy we have 022.0141.0157.0 =× . We will see later, after the BOCR merits are weighted, that the priorities of nine of these (shown in boldface), Military Capability, Technological Advancement, Arms Sales, Spin-Off, Security Threat, Sunk Cost, Further Investment, Arms Race, and Technical Failure account for approximately 0.760 of the total. To economize ef- fort, we used these nine as control criteria each with its decision network to do the analysis. Actually we simply chose the top ones under each merit without being fully consistent about the cutoff point. For example we left out U.S. Reputation under Risks. All economic cost factors were included. We proceeded as if these nine criteria and subcriteria, called covering criteria for the alternatives, were the only criteria to drive the outcome. Their decision networks and connections are shown in Fig. 1–9. A more thorough analysis might include a few more criteria or subcriteria. President/Military Congress Tech. Experts Defense Industry Foreign Countries Allies Alternatives Deploy NMD Termination Global Defense R&D Congress Tech. Experts Industry Military Fig. 1. Decision Network under The Military Capability. Control Subcriterion of Benefits The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 9 T a b l e 1 . Criteria and Their Priorities Merits Criteria Sub-criteria Global Priorities (Normalized) Local Economy (0.141) 0.022 Economic (0.157) Defense Industry (0.859) 0.135 Bargaining Power (0.859) 0.064 Political (0.074) U.S. Military Leadership (0.141) 0.010 Deterrence (0.267) 0.128 Military Capability (0.590) 0.284 Security (0.481) Antiterrorism (0.143) 0.069 Tech. Advancement (0.834) 0.240 Benefits Technology (0.288) Tech. Leadership (0.166) 0.048 Arms Sales (0.520) 0.520 Spin- off (0.326) 0.326 Space Development (0.051) 0.051 Opportunities Protection of Allies (0.103) 0.103 Security Threat: Vulnerability to the security threat 0.687 Sunk Cost (0.539) 0.123 Economic (0.228) Further Investment (0.461) 0.105 ABM Treaty (0.589) 0.050 Costs Political (0.085) Foreign Relations (0.411) 0.035 Technical Failure (0.430) 0.430 Arms Race (0.268) 0.268 Increased Terrorism (0.052) 0.052 Environmental Damage (0.080) 0.080 Risks U.S. Reputation (0.170) 0.170 Tech. Experts Foreign Countries Allies Deploy NMD Alternatives Termination Global Deftnce R&D Congress Congress Tech. Experts Defense Industry Industry President/Military Military Fig. 2. Decision Network under The Technological Advancement. Control Subcriterion of Benefits Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 10 Alternatives Deploy NMD Termination Global Defense R&D Congress Congress President/Military Military Foreign Countries Other Superpowers Adversary Countries Terrorists Allies Tech. Experts Tech. Experts Defense Industry Industry Fig. 3. Decision Network under The Arms Sales. Control Criterion of Opportunities Alternatives Deploy NMD Termination Global Defense R&D Congress Congress Tech. Experts Tech. Experts Defense Industry Industry President/Military Military Fig. 4. Decision Network under The Spin-Off. Control Criterion of Opportunities Congress Congress President/Military Military Foreign Countries Other Superpowers Adversary Countries Terrorists Allies Alternatives Deploy NMD Termination Global Defense R&D Fig. 5. Decision Network under the Security Threat. Control Subcriterion of Costs The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 11 Fig. 7. Decision Network under The Further Investment. Control Subcriterion of Costs Congress Congress President/Military Military Defense Industry Industry Alternatives Deploy NMD Termination Global Defense R&D Tech. Experts Tech. Experts President/MilitaryCongress Alternatives Deploy NMD Termination Global Defense R&D Congress Military Fig. 6. Decision Network under the Sunk Cost. Control Subcriterion of Costs Congress Congress President/Military Military Defense Industry Industry Alternatives Deploy NMD Termination Global Defense R&D Tech. Experts Tech. Experts Fig. 8. Decision Networks Under the Technical Feasibility. Control Criterion of Risks Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 12 3. FULL DEVELOPMENT OF THE ANALYSIS WITH RESPECT TO A SINGLE CRITERION We explain in outline form our thinking about the network under one of the crite- ria. We have chosen Military Capability, one of the main control subcriteria, to elaborate the details of its decision network. There are five main parties involved in the decision making process of NMD: Congress, President/Military, Foreign Countries, Technical Experts and the Defense Industry. The latter two influence Congress and President/Military by providing their professional expertise and technical information. Allies among Foreign Countries can have a partial influ- ence on Global Defense among the four alternatives through economic and tech- nological cooperation. The first block of four rows and four columns in Tabl. 2, a, The Unweighted Supermatrix, indicates that Deploy NMD (NMD) and R&D (R&D) are influenced by Global Defense (Glob~) with priorities of 0.5760 and 0.4240 respectively. The next five columns and first four rows of Tabl. 3, a. The Unweighted Supermatrix, summarize the different views of actors on the contribution of each of the four alternatives to U.S. military capability. Congress, President/Military, Defense In- dustry, and Technical Experts all have a say as to what extent the decision con- tributes to the Military Capability of the U.S. All domestic actors think that De- ploy NMD will increase military capability followed by Global Defense, R&D and Termination (Term~) but to different degrees. Deploy NMD (0.5587) was given the highest priority by Defense Industry, followed by the priority given by President/Military (0.5158), and Congress (0.5060). The lowest priority given to NMD is by Technical Experts (0.2878). It reflects the opinion of scientists who think Deploy NMD is technically infeasible and would not contribute to the en- hancement of U.S. military capability. Only Global Defense is influenced by Al- lies and thus the priority of Global Defense is equal to 1.0000. The fifth to the last row of Tabl. 2,a show connections among components (clusters) each consisting of a single element except for the component of Alter- natives that has four elements. The priorities of the entries in these rows must be either 1.0000 or 0.0000 depending on whether there is influence among them. For Congress Congress President/Military Military Foreign Countries Other Superpowers Adversary Countries Terrorists Allies Alternatives Deploy NMD Termination Global Defense R&D Fig. 9. Decision Network under the Arms Race. Control Criterion of Risks The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 13 example, the fifth to the ninth entries of column one have unit entries obtained from answering the question “Is the component of Congress influenced by Deploy NMD?,” “Is the component of Defense Industry influenced by Deploy NMD?” and similarly for the other three alternatives. All actors are influenced by the three alternatives of Deploy NMD, Global Defense and R&D. Note that an entire col- umn under Termination in the Unweighted Supermatrix of Tabl. 2,a consists of zeros because nothing is influenced by Termination and that leads to dropping the entire matter of missile defense. It is worth noting that under the Security Threat criterion of Costs (not shown here), the column under Termination in the Un- weighted Supermatrix consists of non-zero values because security threat to the U.S. would continue particularly if Termination is chosen as it accentuates vul- nerability of U.S. security. Tabl. 2,b shows the pairwise comparisons of the components. The judgments were obtained by answering the question “Which of two components is influ- enced more by a third component with respect to military capability?” The eigen- vectors of the pairwise comparisons of the components in the matrices of Tabl. 2,b are exhibited in Tabl. 2,c, augmented by zeros in those positions where the components on the left are not influenced by the component on top of the col- umn. The Weighted Supermatrix of Tabl. 2,d illustrates the weighting of the blocks of the supermatrix by the priorities from the corresponding eigenvector of comparisons of the components in Tabl. 2,c. Tabl. 2,e, The Limit Supermatrix, yields the stable priorities of all the elements. From it, the priorities of the four alternatives are extracted and normalized. We obtain for (Deploy NMD, Global Defense, R&D, and Termination) the corresponding values (0.1532, 0.0968, 0.0438, 0.0201) which when normalized by dividing by their sum yields the prior- ity vector (0.488, 0.308, 0.140, and 0.064). This vector is included on the right of the first row of Tabl. 3,4. Similar computations are done for the remaining eight high priority criteria and their normalized results are included in Tabl. 3. An entry in each subcolumn of the supermatrix indicates the relative priority within the block to which that subcolumn belongs that an element on the left is influence by the element on top of the column with respect to Military Capability. Each sub- column is an eigenvector imported from a corresponding pairwise comparisons matrix not shown here because its elements can be approximately formed from the ratios of the corresponding priority vector. A subcolumn of zeros indicates no influence and therefore no comparisons matrix is needed. T a b l e 2 . All Matrices for the Military Capability Decision Network of Benefits T a b l e 2 , a . The Unweighted Supermatrix MilCap Altern~ Cong~ Def. Ind~ For~ Pre/Mil~ Tech~ Unweighted NMD Glob~ R & D Term~ Cong~ Industry Allies Military Tech~ Altern~ NMD 0.0000 0.5760 0.0000 0.0000 0.5060 0.5587 0.0000 0.5158 0.2878 Glob~ 0.0000 0.0000 0.0000 0.0000 0.2890 0.2574 1.0000 0.2929 0.2623 R & D 0.0000 0.4240 0.0000 0.0000 0.1307 0.1382 0.0000 0.1367 0.2369 Term~ 0.0000 0.0000 0.0000 0.0000 0.0744 0.0457 0.0000 0.0546 0.2130 Cong~ Cong~ 1.0000 1.0000 1.0000 0.0000 0.0000 1.0000 1.0000 1.0000 1.0000 Defense Ind~ Industry 1.0000 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 For~ Allies 1.0000 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 Pre/Mil~ Military 1.0000 1.0000 1.0000 0.0000 1.0000 1.0000 1.0000 0.0000 1.0000 Tech~ Tech~ 1.0000 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 14 T a b l e 2 , b . Pairwise Comparisons Matrices and Priorities of Components Pairwise comparing components with respect to the Alternatives component Q: Which of a pair of components is influenced more by the Alternatives component with respect to Military Capability? Altern~ Cong~ Def. Ind~ For~ Pres~ Tech~ Prior. Altern~ 1.0000 0.1667 0.2500 1.3300 0.1429 0.5556 0.0486 Cong~ 5.9999 1.0000 2.2000 6.2000 0.7407 3.2000 0.2889 Def. Ind~ 4.0000 0.4546 1.0000 4.0000 0.4115 2.2600 0.1653 For~ 0.7519 0.1613 0.2500 1.0000 0.1250 0.5263 0.0425 Pres~ 7.0000 1.3500 2.4300 8.0000 1.0000 5.1000 0.3742 Tech~ 1.8000 0.3125 0.4425 1.9000 0.1961 1.0000 0.0805 Pairwise comparing components with respect to the Congress component Pairwise comparing components with respect to the Defense Industry component Q: Which of a pair of components is influenced more by the Congress component with respect to Military Capability? Q: Which of a pair of components is influ- enced more by the Defense Industry compo- nent with respect to Military Capability? Altern~ Pres~ Prior. Altern~ Cong~ Pres~ Prior. Altern~ 1.0000 0.5638 0.3605 Altern~ 1.0000 0.6769 0.5388 0.2292 Pres~ 1.7736 1.0000 0.6395 Congr~ 1.4773 1.0000 0.6600 0.3181 Pres~ 1.8561 1.5152 1.0000 0.4528 Pairwise comparing components with respect to the Foreign Countries component Pairwise comparing components with respect to the Presidnet/Military component Q: Which of a pair of components is influ- enced more by the Foreign Countries com- ponent with respect to Military Capability? Q: Which of a pair of components is influ- enced more by the President / Military com- ponent with respect to Military Capability? Altern~ Cong~ Pres~ Prior. Altern~ Cong~ For~ Prior. Altern~ 1.0000 0.5556 0.3259 0.1671 Altern~ 1.0000 2.1887 3.6604 0.5735 Cong~ 1.8000 1.0000 0.4632 0.2781 Congr~ 0.4569 1.0000 2.0377 0.2799 Pres~ 3.0682 2.1591 1.0000 0.5548 For~ 0.2732 0.4907 1.0000 0.1467 Pairwise comparing components with respect to the Technical Experts component Q: Which of a pair of components is influenced more by the Technical Ex- perts component with respect to Military Capability? Altern~ Cong~ Pres~ Prior. Altern~ 1.0000 2.5379 2.5379 0.5593 Congr~ 0.3940 1.0000 1.0000 0.2204 Pres~ 0.3940 1.0000 1.0000 0.2204 T a b l e 2 , c . Priorities Matrix of Eigenvectors How much components are influenced by each component; imported from the matrices of Tabl. 3b above Clusters Altern~ Cong~ Def. Ind~ For~ Pres~ Tech~ Altern~ 0.0486 0.3605 0.2292 0.1671 0.5735 0.5593 Cong~ 0.2889 0.0000 0.3181 0.2781 0.2799 0.2204 Def. Ind~ 0.1653 0.0000 0.0000 0.0000 0.0000 0.0000 For~ 0.0425 0.0000 0.0000 0.0000 0.1467 0.0000 Pres~ 0.3742 0.6395 0.4528 0.5548 0.0000 0.2204 Tech~ 0.0805 0.0000 0.0000 0.0000 0.0000 0.0000 The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 15 T a b l e 2 , d . The Weighted Supermatrix Priorities from Tabl. 3c are used to weight corresponding blocks of unweighted supermatrix of Tabl. 3,a MilCap Altern~ Cong~ Def. Ind~ For~ Pre/Mil~ Tech~ Weighted NMD Glob~ R & D Term~ Cong~ Industry Allies Military Tech~ Altern~ NMD 0.0000 0.0280 0.0000 0.0000 0.1824 0.1280 0.0000 0.2958 0.1610 Glob~ 0.0000 0.0000 0.0000 0.0000 0.1042 0.0590 0.1671 0.1680 0.1467 R & D 0.0000 0.0206 0.0000 0.0000 0.0471 0.0317 0.0000 0.0784 0.1325 Term~ 0.0000 0.0000 0.0000 0.0000 0.0268 0.0105 0.0000 0.0313 0.1191 Cong~ Cong~ 0.3037 0.2889 0.3037 0.0000 0.0000 0.3181 0.2780 0.2799 0.2204 Defense Ind~ Industry 0.1737 0.1653 0.1737 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 For~ Allies 0.0446 0.0425 0.0446 0.0000 0.0000 0.0000 0.0000 0.1467 0.0000 Pre/Mil~ Military 0.3933 0.3742 0.3933 0.0000 0.6395 0.4528 0.5548 0.0000 0.2204 Tech~ Tech~ 0.0846 0.0805 0.0846 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 T a b l e 2 , e . The Limit Supermatrix The weighted supermatrix raised to sufficiently large powers to stabilize within rounded off four place decimals MilCap Altern~ Cong~ Def. Ind~ For~ Pre/Mil~ Tech~ Limited NMD Glob~ R & D Term~ Cong~ Industry Allies Military Tech~ Altern~ NMD 0.1532 0.1532 0.1532 0.0000 0.1532 0.1532 0.1532 0.1532 0.1532 Glob~ 0.0968 0.0968 0.0968 0.0000 0.0968 0.0968 0.0968 0.0968 0.0968 R & D 0.0438 0.0438 0.0438 0.0000 0.0438 0.0438 0.0438 0.0438 0.0438 Term~ 0.0201 0.0201 0.0201 0.0000 0.0201 0.0201 0.0201 0.0201 0.0201 Cong~ Cong~ 0.2224 0.2224 0.2224 0.0000 0.2224 0.2224 0.2224 0.2224 0.2224 Defense Ind~ Industry 0.0513 0.0513 0.0513 0.0000 0.0513 0.0513 0.0513 0.0513 0.0513 For~ Allies 0.0619 0.0619 0.0619 0.0000 0.0619 0.0619 0.0619 0.0619 0.0619 Pre/Mil~ Military 0.3255 0.3255 0.3255 0.0000 0.3255 0.3255 0.3255 0.3255 0.3255 Tech~ Tech~ 0.0250 0.0250 0.0250 0.0000 0.0250 0.0250 0.0250 0.0250 0.0250 4. BOCR WEIGHT DEVELOPMENT The judgments used in this analysis were our interpretation of what experts thought about the various issues obtained from the vast reading of the literature we examined and from following the news closely for a period of more than six months. We also consulted some knowledgeable people on the subject in the area. We quickly realized there is no single expert in all the criteria we considered. Sensitivity analysis given later would essentially vary these judgments widely to determine the stability of the outcome. The assessment criteria used to determine the priorities of the BOCR merits are shown in Fig. 10. These are World Peace, Human Well-being, and International Politics. All these criteria have subcriteria under them. The three subcriteria, Adversary Countries, Security Dilemma and Terrorism cover all the causes disturbing or stabilizing peace in the world. The first subcriterion, Adversary Countries, concerns the potential threats by adver- sary countries. The second criterion, Security Dilemma, means that increasing one country’s security inevitably decreases other countries’ security. Terrorism indi- Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 16 cates any possibility of the rise or decline of terrorism in the world. Human Well- being includes Technological Advancement and Market Creation. Technological Advancement driven by the NMD research and development process can ulti- mately benefit all people, particularly in providing possible space exploration that can lead to the creation of new markets. Moreover, the 21st century is character- ized as a post-industrialization era. Service industries in communication and transportation will benefit not only businesses associated with these industries, but also consumers who can enjoy the products from the new market. The last criterion is International Politics. It is composed of two subcriteria, Military Rela- tions and Diplomatic Relations. Military Relations refer to the impact of NMD on relations with U.S. allies for better or for worse. Also, the impact of NMD on dip- lomatic relations among all countries should be considered. The priorities shown next to the criteria and subcriteria in Fig. 10 were obtained through the usual pairwise comparison process of the AHP according to their importance with re- spect to their higher-level goal or parent criterion. The four merits of BOCR were rated according to five intensities listed be- low along with their priorities. The outcome is summarized in Tabl. 3. The inten- sities were derived from pairwise comparisons. T a b l e 3 . Priority Ratings for the Merits: Benefits, Opportunities, Costs and Risks Very High (0.419), High (0.263), Medium (0.160), Low (0.097), Very Low (0.061) Benefits Opportunities Costs Risks Adversary Countries Very High Medium High Very Low Security Dilemma Very Low Very Low Very High Very Low World Peace Terrorism Medium Very Low High High Technological Ad- vancement High High Low Very Low Human Well-Being Market Creation Medium High Very Low Very Low Military Relations High High Medium Very Low Interna- tional Poli- tics Diplomatic Relations Low Low Low Very High Priorities 0.264 0.185 0.361 0.190 Note that BOCR are rated one at a time and are not obtained from paired comparisons. They are obtained using the rating approach of the AHP. As we said earlier if we weight the priorities derived in Tabl. 1 by the corre- sponding priorities of the merits just derived and then add we get: EVALUATING MERITS WORLD PEACE Adversary Countries Security Dilemma Terrorism HUMAN WELL-BEING Technological Advancement Market Creation INTERNATIONAL POLITICS Military Relations Diplomatic Relations Fig. 10. Strategic Criteria for BOCR Ratings The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 17 +×+×+×+×+× 687.0361.0326.0185.0520.0185.0240.0264.0284.0264.0 76.0268.0190.0430190.0105.0361.0123.0361.0 ≈×+×+×+×+ . In most of our studies we attempt to use factors whose total weight in not less than 70%. 5. OVERALL OUTCOME Tabl. 4 shows the priorities of the nine control criteria or subcriteria, the corre- sponding priorities of the alternatives that are normalized from Tabl. 2,e, The Limit Supermatrix, their synthesis for each of the BOCR merits together with the normalized reciprocals under costs and risks. The final outcome in Tabl. 5 is de- rived by weighting the synthesized priorities of the alternatives of Tabl. 4 by the priorities of the BOCR merits, again using the reciprocals of the synthesized pri- orities of the alternatives under costs and risks. T a b l e 4 . Synthesized Priorities of the Nine Control Criteria and Subcriteria Merits Criteria Subcriteria Deploy NMD Global Defense R&D Termination Benefits (0.264) Security (0.481) Military Capa- bility (0.590) 0.488 0.308 0.140 0.064 Technical (0.288) Technical Ad- vancement (0.834) 0.364 0.398 0.172 0.066 Benefits Synthesized 0.226 0.183 0.081 0.034 Benefits Normalized 0.431 0.349 0.155 0.065 Arms Sales (0.520) 0.483 0.300 0.145 0.072 Opportunities (0.185) Spin-Off (0.326) 0.506 0.264 0.146 0.084 Opportunities Synthesized 0.416 0.242 0.123 0.065 Opportunities Normalized 0.492 0.286 0.145 0.077 Costs (0.361) Security Threat (0.687) 0.087 0.164 0.275 0.475 Economic (0.228) Sunk Cost (0.539) 0.476 0.273 0.158 0.092 Further In- vestment (0.461) 0.525 0.258 0.143 0.074 Costs Synthesized 0.173 0.173 0.223 0.345 Costs Normalized 0.189 0.189 0.244 0.377 Costs Reciprocal 0.305 0.305 0.236 0.153 Technical Failure (0.430) 0.473 0.269 0.154 0.103 Risks (0.190) Arms Race (0.268) 0.410 0.284 0.181 0.124 Risks Synthesized 0.313 0.192 0.115 0.078 Risks Normalized 0.448 0.275 0.165 0.112 Risks Reciprocal 0.107 0.174 0.291 0.428 Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 18 T a b l e 5 . Overall Results Alternatives Benefits Opportuni- ties Costs Risks Overall 0.264 0.185 0.361 0.190 Deploy NMD 0.431 0.492 0.305 0.107 0.335 Global Defense 0.349 0.286 0.305 0.174 0.288 R & D 0.155 0.145 0.236 0.291 0.208 Termination 0.065 0.077 0.153 0.428 0.168 Deploy NMD (0.335) scores the highest. It is a comprehensive result that takes into consideration all BOCR. The conclusion of this analysis is that pursu- ing the deployment of NMD is the best alternative. This is because, as it is shown in Tabl. 5, Deploy NMD has the highest priorities for three (benefits, opportuni- ties and costs) of the four merits. But we must now examine how realistic this outcome is. 6. SENSITIVITY ANALYSIS One might have different judgments in comparing the importance of BOCR or of the nine control criteria. To ensure the stability of the outcome of our analysis, we conducted sensitivity analysis. A. Sensitivity Analysis at the BOCR Level First, we increased and decreased one of the four merits of BOCR keeping the others proportionally the same. For example, if benefits were to be increased from its original priority 0.264 to 0.500, the sum of the other three merits would com- prise the other 0.500 and the proportion among them would remain the same as before and their new priorities would be: opportunities, 0.124, costs, 0.246, and risks, 0.130. We found that no matter how much we increased or decreased the priorities of benefits, opportunities and costs the overall ranks of the final out- come were preserved although these experiments changed the magnitude of the superiority of the best alternative, Deploy NMD (for example, from 0.301 to 0.431 for benefits as Fig. 11 shows). Only changing the priority of risks reversed the ranks of the four alternatives. This occurred only when the priority of the risks were as large as 0.375 or more. Then, Termination gradually became third then second and finally the best alternative as the priority of risks was increased more and more (Figs 12). B. Sensitivity Analysis at the Control Criterion Level We did similar tests for the nine criteria that have decision networks. We found that the outcome was very stable and did not change the overall ranks except for changes of the three criteria: Security Threat, Sunk Cost and Further Investment all under costs. When the priority of Security Threat decreased to about 0.172 from 0.687 (Fig. 13) or the priority of Sunk Cost increased to 0.753 (Fig. 14) or the priority of Further Investment increased to 0.734 (Fig. 15), Termination gradually began to move to third, second and finally to first rank position. The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 19 Some are highly concerned with risks associated with NMD, such as Tech- nical Failure and Arms Race. We did another test using larger priorities for risks to see if it would change the outcome. In that case, the control criterion, U.S. Reputation, under risks replaced the control criterion, Further Investment, under costs. Interestingly enough, the ranks of the alternatives were the same as in Tabl. 5 with a slightly higher priority for Deploy NMD. Our sensitivity analysis indicates that the final ranks of the alternatives might change, but such change requires making extreme assumptions on the priorities of BOCR and of their corresponding control criteria. The outcome in Tabl. 5 is very stable and the United States should choose Deploy NMD as the best al- ternative for the decision. I am grateful to my colleague Professor Dr. Klaus Dellmann for his careful reading and suggestions to improve the paper and to my student Yeonmin Cho for her untiring efforts in preparing the case study and above all Mujgan Ozdemir for helping me prepare this final version. Fig. 12. Sensitivity Analysis for Risks. Termination becomes the more preferred alterna- tive as the priority of risks increases Fig. 11. Sensitivity Analysis for Benefits. The rank remains the same regardless of the priorities of benefits Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 20 Fig. 13. Sensitivity Analysis for Security Threat: If the priority of Security Threat be- comes less than about 0.172, Termination becomes the more preferred alternative Fig. 15. Sensitivity Analysis for Further Investment. If the priority of Further Investment becomes larger than 0.734, Termination becomes the more preferred alternative Fig. 14. Sensitivity Analysis for Sunk Cost. If the priority of Sunk Cost becomes larger than about 0.753, Termination becomes the more preferred alternative The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 21 7. US POLICY TOWARDS IRAQ For this example done jointly with my student Bethany Simunich in October 2002, we give the Tabl. 6 and Figs 16–20 without as much discussion as in the previous example. This is a two level model with the decision subnets directly attached to the merits and with no control criteria. EVALUATING MERITS ECONOMIC Monetary Cost of War Change in Oil Prices Focus Shift away from U.S. Economy MILITARY WMD U. S. Military casualties Removal of Dictator POLITICAL Allied reaction Arab world reaction Regime change SOCIAL Iraqi civilians Public Opinion Fig. 16. Strategic Criteria for Rating the BOCR Merits Fig. 17. The Decision Subnet under Benefits Thomas L. Saaty ISSN 1681–6048 System Research & Information Technologies, 2003, № 4 22 Рис. 19. The Decision Subnet under Opportunities Fig. 18. The Decision Subnet under Costs Fig. 20. The Decision Subnet under Risks The analytic network process. Examples. Part 2.3 Системні дослідження та інформаційні технології, 2003, № 4 23 T a b l e . 6 . Overall Outcome Alternatives Priorities Pre-emptive Attack on Iraq 0.17 Attack Iraq only with Allied Help 0.27 Work with UN to Ensure Weapons Inspections 0.37 Remove Sanctions 0.19 It appears at this writing, the war has just begun, that it would have been bet- ter for USA to work with the U.N. although by combining the priorities of the first two alternatives, one may infer that the course followed in practice is also justified by this analysis. REFERENCES 1. Saaty, Thomas L. The Analytic Network Process, RWS Publications, 4922, Ellsworth Avenue, Pittsburgh, Pa. 15213. 2. Saaty, Thomas L. Theory Analytic Hierarchy and Analytic Network Processes- Examples, Part 2.2. // System research and information technologies. — № 2. — 2003. — P. 7–33. 3. Simunich, B. An Evaluation of U.S. Foreign Policy re: Iraq Using the ANP, October 2002, Unpublished Project. Received 03.09.2003 From the Editorial Board: The article corresponds completely to submitted manuscript.

The Analytic Network Process. Examples. Part 2.3

Institution

Ähnliche Einträge