Application of the Co-Evolutionary Algorithm with Memory at the Population Level for Optimisation of the Operation of Real Electric Power Distribution Networks

Abstract

A solution of the optimisation problem of complex power electric networks post-fault configuration has been proposed. The analysed problem of search for optimal configurations of electric power distribution networks for changing loadings of network elements and for malfunction conditions is a multi-criteria optimisation problem. In this case the sought-after solution is the collection of Pareto-optimal solutions. Scientific methods belonging to the class of artificial intelligence methods (evolutionary algorithms and classifier system) have been used in the paper. Scientific work of the author is presented the co-evolutionary algorithm using memory at the level of organised populations in the form of five subpopulations, the composition of which is changed and organised according to classifying systems’ procedures. The process of creating a collection of classifiers describing the substitute network configuration was performed by the author supported by the theoretical genetic basics of self-teaching system. Cooperation of the evolutionary algorithm with the classifier system enables significant reduction of the classification time, reduces the iterative calculation process on average by 40 %. The calculations performed for the mapped real system of the medium voltage municipal distribution network have given satisfactory results, confirming the adequate direction of the research. The method presented in the article enables effective search of optimal configurations of distribution networks for various network loadings and also network malfunction conditions