A memetic algorithm for multi-objective resource allocation problems
Angela H. L. Chen, Chiuh‐Cheng Chyu · Journal of Statistics and Management Systems · 2011
This paper presents one application of the multi-objective resource allocation problem, in which decision makers need to allocate between limited resources for the performance of tasks so that the best overall work efficiency is accomplished at the lowest labor cost. With this problem, we apply the memetic algorithm to obtain a set of Pareto optimal solutions. The search procedure consists of three main steps selection, crossover, and mutation. In addition, once the new solutions are generated, a local search scheme will be employed to help exploit the search space. The proposed memetic algorithm is tested on two benchmark instances and compared with two other methods –hGA and ACO in the literature. Three performance metrics are considered – (1) the hit ratio; (2) the accuracy ratio; (3) the D1 R value. Computational results show that, with test instance 1, our MA outperforms hGA and ACO in terms of all three performance measures. While, in test instance 2, our MA is compared with the only available method –hGA from literature. The results of MA show its superiority in all categories, even with a larger instance. Decision makers can be benefited from a set of abundant alternatives provided by the effective and efficient MA.