Decision-Making with Imperfect Information

Alan M Stanier

Thesis submitted in fulfillment of the requirements for the degree of Ph D
Essex University, 1977

Summary

This thesis describes the application of a particular Artificial Intelligence methodology (that of Intelligent Databases) to the task of making rational decisions in a environment of imperfect imformation. The task of bidding at Contract Bridge is used as a paradigm for such situations, it being shown that the more usual games used for such purposes are inadequate models of the Real World.

The methodology is described, together with our reasons for believing the methodology to be superior to any others currently in use.

The results of applying the methodology to the chosen domain are described, and - by the use of illustrative examples - the resulting programs are presented in detail. The results are compared with those obtained by previous programs (including one by the author), and it is shown that this comparison confirms our view of the superiority of Intelligent Database methodologies.

Finally, our proposed future work is described, as are our reasons for believing that such work will lead to a better program, and to a clearer understanding of the Decision-Making Process.

Table of Contents