Applying Memetic Algorithm in Multi-Objective Resource Allocation among Competing Projects

This paper presents a memetic algorithm for solving project resource allocation problems, where the resource requirements of each project concern numbers of monetary units and never exceeds the amount of capital available. Our objectives are to obtain the best overall result for which returns are maximized and costs are minimized. Such problem, considered as a multi-objective optimization, is too complex to be solved by exact methods. In the proposed MA, the population generated by the crossover or the mutation operator is further improved by a local search method. The approximated Pareto front is updated using all new solutions generated. The performance of MA is demonstrated via an instance, consisting of six projects, and 120 units of capitals, and compared with the reference Pareto front found by executing the exhaustive method. Based on both solution quality and CPU time, the results of such comparison have proved our MA to be an effective optimal method in the multi-objective resource allocation problems (MORAP).