Aspiration Based Decision Support Systems

Abstract

IntroductionIt is not easy to summarize -even in a volume -the results of a scientific study conducted by circa 30 researchers, in four different research institutions, though cooperating between them and jointly with the International Institute for Applied Systems Analysis, but working part-time, sponsored not only by IIASA's national currency funds, but also by several other research grants in Poland. The aims of this cooperative study were defined broadly by its title Theory, Software and Testing Examples for Decision Support Systems. The focusing theme was the methodology of decision analysis and support related to the principle of reference point optimization (developed by the editors of this volume and called also variously: aspiration-led decision support, quasi-satisfying framework of rationality, DIDAS methodology etc.). This focusing theme motivated extensive theoretical research -from basic methodological issues of decision analysis, through various results in mathematical programming (in the fields of large scale and stochastic optimization, nondifferentiable optimization, cooperative game theory) motivated and needed because of this theme, through methodological issues related to software development to issues resulting from testing and applications. We could not include in this volume all papers -theoretical, methodological, applied, software manuals and documentation -written during this cooperative study. The selection principle applied for this volume was to concentrate on advances of theory and methodology, related to the focusing theme, to supplement them by experiences and methodological advances gained through wide applications and tests in one particular application area -the programming of development of industrial structures in chemical industry, and finally to give a very: short description of various software products developed in the contracted study agreement. The material of this volume is thus divided correspondingly into three unequal parts (it must be noted, however, that the last and shortest part corresponds to the most extensive research effort).PART 1 is composed of 15 theoretical and methodological papers. It starts with two more general papers, the first explaining the focusing theme of this volume and the second describing the methodology of decision analysis in decision support systems (DSS) of the DIDAS family. The following five papers are devoted to various aspects of linear programming: three represent innovative approaches to large-scale programming problems and new mathematical and algorithmic results in this field, including a new idea of decomposition of augmented Lagrangian functions for large-scale problem -motivated and related to the work on DSS, next addresses basic problems of multiobjective dynamic trajectory optimization, a further one presents a more detailed methodological guide to a multiobjective mathematical programming package HYBRID. We present such a mixture of results on purpose, to show the broad scope of the study, its com-IV ponents of mathemati...