Knowledge guided genetic algorithm for optimal contracting strategy in a typical standing reserve market

This paper proposes a knowledge guided genetic algorithm (GA) when used for searching for the optimal contracting strategy in a typical standing reserve market. The knowledge is effectively used for significantly reducing search space. The knowledge is obtained by identifying subset of tenders that have similar contract patterns among low cost solutions and relative large impact on the final solution results. Once the subset is detected, it is then fixed throughout the subsequent GA searches so that the GA can work in a much reduced problem space and concentrate on areas that need most attentions. This search space reduction has demonstrated on a system with 83 tenders. The simulation results clearly show that the search space reduction has significantly improved final solution cost and solution robustness when comes to meet operating reserve requirements.