THREE SUBJECTS OF THE ADAPTIVE DESIGN AND ANALYSIS FOR MULTI-STAGE SIMULATION SYSTEM.

This paper describes three main results which have been developed through one of the Waseda-IBM Japan joint research project, entitled 'Adaptive Design and Analysis for Multistage Simulation Systems (ADAMS)'. The first part develops a method to enable a new generation computer to reason and/or search in a close way to the process of human problem-solving. It considers the cases in which M types of information are simultaneously transmitted to one node or one type of information is transmitted to M diverging points. It shows that M dimensional CUBE can represent these cases and M dimensional CUBE can be described by means of the ternary system. The paper considers several forecasting models which seem to be frequently used in multi-stage simulation. It focuses on evaluation criteria for forecasting precision, forecasting errors and testing methods. It presents a conceptual framework for an intelligent expert system for decision-supporting in economics and management.