Knowledge Synthesis on Multidimensional Matrices

The tables containing the optimal decisions obtained when solving real decision-making problems under uncertainty are often extremely large. This raises problems related not only to the storage and management of so much information, but also to the use of these tables for knowledge retrieval and reasoning explanation purposes. In this paper, we propose turning the tables into minimum storage multidimensional matrices. Computers manage multidimensional matrices as lists, where each position is a function of the order chosen (or base) for the matrix dimensions. The optimal list includes the same knowledge as the original list, but it is compacted, which is very valuable for explaining expert reasoning. Thus, evolutionary computation is required to minimise the number of list entries. The process needs a learning procedure because of its complexity. We illustrate the ideas using our decision support system IctNeo [1] for neonatal management, outputting excellent results. The methodology is so general that it also applies to any table considered as a knowledge base. Keywords— Decision Support Systems, Decision Analysis, Influence Diagram, Decision Table, Multidimensional Matrix, KBM2L List, Combinatorial Optimisation, Explanation