Abstract:To the problem of dimensionality when the large scale unit commitment problems are solved by the traditional forward dynamic programming method,the storage of local optimizations and the small probability mutation mechanism are adopted to select the path storage in dynamic programming,which reduces the calculation of the path evaluation and also improves the exploration ability of the algorithm.In order to avoid the local extreme value of the modified algorithm,the tabu list is set up to prevent the repeated path searching.A constructing method which compresses state space based on trial stop optimization is proposed,which reduces the size of initial state space and adjusts the neighborhood path exploration method for the problem of over compression.The simulation results of 10-60 units verify the feasibility of the proposed algorithm.