Abstract
Power resource planning-the decision of construction period of generators-is affected by several uncertainties: load demand, fuel cost, and so on. In order to evaluate power resource planning which necessarily involves several uncertainties, it is usual for planners to prepare many scenarios which consider several uncertainties and to analyze whether the particular resource plan is feasible far every scenario. It is a large computational burden to analyze even one scenario. In order to realize an efficient scenario analysis, several preferable scenarios which can reflect many scenario must be determined. In this paper, in order to create preferable scenarios, we propose a new scenario creation algorithm corresponding to the planner's consideration process. The proposed method employs a combntation of Chaotic Neural Network (CNN) and Genetic Algorithm (GA). CNN is able to carry out a dynamic pattern association function effectively, and GA has an excellent adaptive nature. The effectiveness of the proposed method is demonstrated by numerical results for the actual data of several uncertainties in 1986_??_1990 through comparing these with the decision tree method.
