While ANNs experienced various difficulties during their development, it was found that the integration of genetic algorithms (GA) could offer a better rate of success (Ku§çu & Thornton, 1994; Mitchell, 1996). The study of GA was initiated by John Holland in the 1970s (Tsoukakas & Uhrig, 1997) based on the principle of Darwin’s the survival of the fittest. GA is also useful to many other
5 A n inform al definition o f b est approxim ation was given by G irosi & Poggio (1989), “A n app ro x im atio n schem e has the b e s t ap p roxim ation property if in th e set o f a p p ro x im atin g functions (for instance the set F(W,x) spanned by param eters W) there is one th a t has a m inim um distance fro m any given fu n ctio n o f a larger set O .”
disciplines; for example, optimisation techniques (Tsoukakas & Uhrig, 1997, Ignizio, 1998), multi-disciplinary design optimisation (MDO) (Coit & Smith, 1996), and reliability design (Monga & Zuo, 1998). The military applications for GA include the helicopter conceptual design (Crossley & Laananen, 1997) and the prediction of foreign missile effectiveness (WoUam, 1999). Although GA has been applied successfully as a complimentary to ANN and other disciplines, its employment requires the intrinsic characteristics of object, and hence will not be applied to the ANN approach of this dissertation.
2.5 Summary
The complexities of weapon system evaluation have driven the search for a simpler or visual decision process. Although many advanced techniques have been developed, most of them are applicable to linear problems, although some non-Hnear method development are undergoing. Traditional MCDM methods such as AHP have been applied in various attempts to evaluate alternative weapon systems due to its systematic and hierarchic approach; however, the inconsistencies of human judgement involved in the analysis process have been concerned and modified by many analysts. It has been shown that the application of artificial intelligence has drawn a great attention in overcoming the inconsistencies involved AHP and other methods. The proposal of value-focused thinking (VFT) by Keeney (1992) provides a sound framework for evaluating complex systems focusing on their intrinsic properties; however, in the real world, the evaluation of alternatives can only be implemented through their attributes, which have to take-as-they-are, especially for the aircraft selection problems to be discussed in this study.
The development of UTA Qacquet-Lagreze & Siskos, 1982) is opposite to most
MCDM approaches; however, it is identical to the scenarios as suggested in this smdy. How to capture a decision maker’s preference profile and repeat it when
evaluations of alternatives need to be implemented. UTA solves the MCDM problems with ranking order, while this smdy intends to predict their exact values. ANN has been attempted by others to solve MCDM problems in terms of ranking orders, it has shown promising in bridging the need of evaluate alternatives by value-focused methodologies.
CHAPTER 3
COMBAT AIRCRAFT ACQUISITION ISSUES
Combat aircraft is selected to represent the complexities in weapon systems acquisition and the issues involved in judging alternative systems’ value in this dissertation. First-tier arms-producing countries focus their weapon systems acquisitions on designing systems according to operational requirements. The complexities of weapon systems are dealt with through well-developed systems engineering process and optimisation techniques. In the second-tier arms- producing countries, the Republic of China is an example ^ ; major weapon systems such as combat aircraft mainly rely on off-the-shelf procurement. Off- the-shelf procurement is perceived as a process of finding solutions that satisfy requirements, which is an inverse of the systems engineering process employed by first-tier arms producers. Compared with the approach of new system development, the techniques applied to off-the-shelf procurement were less discussed. However, effectiveness evaluations of alternatives are critical in the early phase of weapon systems acquisition.
This chapter wül reveal potential complexities involved in an off-the-shelf procurement by reviewing weapon system acquisition practices and the systems engineering framework. Warship design wül be used to iUustrate the complex inter-relationships of weapon systems. The analyses of combat aircraft
■* U nited States, the U n ited K ingdom , France, G erm any, and Italy (Bitzinger, 2000)
5 A ccording to Bitzinger, the “ third-tier” states define th o se countries w ith low -tech arm s-p ro d u c tio n capability, for exam ple, Egypt, Mexico, and Nigeria. T h e “ second-tier” countries are u sed to include th o se countries b etw een the first-tier and third-tier countries; for exam ple, A ustralia, Ja p an , Sw eden, Brazil, Iran, Singapore, S outh Africa, South K orea, Taiwan, Turkey, C hina, and India (ibid.).
performance and effectiveness are used to explain the possible shape of complex intra-relationships and build the objective tree for decision analysis.
The rest of the chapter are organised as follows. Section 3.1 reviews the overall environment of weapon systems acquisition. Section 3.2 explains the systems engineering process applied in the development of weapon systems. Section 3.3 describes typical criteria used to evaluate weapon systems. Section 3.4 explains the attributes employed to characterise the performance of combat aircraft. Section 3.5 establishes the value tree of combat aircraft and describes the data of specification gathered from various sources. Section 3.6 summarises the issues of this chapter.