Abstract: Column generation is employed to solve the multi-project scheduling problem in this paper. Several constraints,such as human resources'capability and level,are considered in this model. In addition, we consider the problem of how to minimize the total costs,including fixed and variable,while satisfying the given constraints. The model is decomposed and solved according to column generation procedure. Due to its complexity,a heuristic algorithm is developed to solve the corresponding sub-problems:The priority rule based heuristic algorithm provides the initial solutions first;then,the immune genetic algorithm is introduced to obtain further improvements The experiment results show the algorithm performance,the model improvement and the influence of the different priority rule combinations on the objective and time allocation between projects.

Key words: project scheduling, human resource, multiple project, column generation, genetic algorithm