Abstract: A twostage optimization algorithm was represented to solve one type of project scheduling problems with multiskilled workforce constraints(PSPMSWC). The model of PSPMSWC was proposed and decomposed into submodels corresponding to the algorithm. By using hybrid genetic local search(GLS) and CPLEX optimization software， a strategy to iteratively solve the submodels in sequence was proposed to ensure that all the instances could be solved successfully. Instances were randomly generated from a standard problem library and utilized to test the performance of algorithm experimentally. Computational results show that all of instances can be solved efficiently， with some of instances having the optimal values. Therefore， the twostage algorithm is proved to be a good method to solve PSPMSWC with practical significance.

Key words: project scheduling, multiskilled workforce, twostage method, genetic local search, assignment problem