Abstract
Stochastic resource allocation problem (SRAP) has been applied in manufacturing system and service system. Resource allocation is a crucial step in system because system performance and cost can be improved by properly allocating resource to each activity. Allocating resource to each activity under resource constraints and minimizing overall costs are the goal of SRAP. However, the system performance can’t be formulated as a closed-form expressions function and system performance is uncertainly. In this study, we proposed a hybrid algorithm based on Simplified Swarm Optimization (SSO) combining with Optimal Computing Budget Allocation (OCBA) and a local search method for SRAP. SSO is applied for global search and OCBA is used to allocate the simulation replications for recognizing the best solution in particle. The core concept of OCBA is allocating the simulation budget to the solutions which has larger standard deviation or superior system performance. Therefore, we can avoid to wasting simulation budget on the solutions with worse system performance and providing reliable solutions. Afterwards, we proposed a local search strategy based on the utilization of machines. By using this local search method we can search local optimum. Finally, the experimental results show that the proposed Hybrid Simplified Swarm Optimization is better than other algorithms.