This most fascinating issue of our journal is primarily the result of Prof. Thomas Sargent, Stanford University, granting permission for the organization of a special session on 'Artificial intelligence in economics and management science' during the 1990 Conference of the Society of Economic Dynamics and Control (SEDC), held in Minneapolis, MN, 28-30 June~ 1990. This issue is also the result of the momentum generated, in that it contains some papers not presented at the SEDC conference, but inspired by it.
We are quite fortunate to see economic and financial applications developed for both neural processing, for expert systems, for constraint logic programming, as well as for qualitative reasoning (static and dynamic). This reflects the fact that researchers from economics and management science are at the forefront of testing out, and contributing to, some new fields of work in artificial intelligence and cognitive science. This helps us all better to compare these techniques with more established approaches, and eventually to integrate the new results into parts of such more traditional solutions.
In their paper, F.S. Wong, P.Z. Wang and H.H. Teh show how to combine neural learning algorithms with uncertain fuzzy rules, for the purpose of learning and reasoning about stock selection strategies in financial markets described by quantitative data in ever-changing contexts. Still another approach to uncertain economic knowledge and data is proposed by E. Krowitz, who essentially discovers that a diagnostic expert system, through its classification of balance of payments situations, is suitable for the analysis of a country's creditworthiness.
The decision support section is covered by J.M. Broek and H.A.M. Daniels in their study of the possible use of constraint logic programming for asset and liability management in banks, by adding numerical and other constraints to logic predicates.
Qualitative reasoning in economics aims at emphasizing the trends in changes of economic variables, and their consistency, in order to predict such trends in agreement with some quantitative economic theory. The paper by K-P. Lin and A.M. Farley deals with one of the more novel uses in disequilibrium conditions where numerical simulation and models alone do not suffice. Extending into how to build such qualitative economic theories, C. Gianotti gives an efficient technique to describe the causalities as perceived by the economic agents and/or model-builders, and how to maintain consistency through a graph propagation algorithm amongst alternatives.