The 0-1 knapsack problem (KP) is a classic NP-hard problem and could be handled by swarm intelligence algorithms. However, most of these algorithms might be trapped in the local optima as the scale increases. Hybrid r...
详细信息
The 0-1 knapsack problem (KP) is a classic NP-hard problem and could be handled by swarm intelligence algorithms. However, most of these algorithms might be trapped in the local optima as the scale increases. Hybrid rice optimization (HRO) is a novel swarm intelligence algorithm inspired by the breeding process of Chinese three-line hybrid rice, its population is classified into three types such as the maintainer, restorer and sterile line and several stages including hybridization, selfing and renewal are implemented. In this paper, a modified HRO algorithm is proposed for the complicated large-scale 0-1 KP. A dynamic step is introduced in the renewal stage to balance the exploration and exploitation phases. Moreover, HRO is combined with binaryantcolonyoptimization (BACO) algorithm to compose the parallel model and serial model for enhancing the convergence speed and search efficiency. In the parallel model, HRO and BACO are independently implemented on two subpopulations and communicate during each iteration. In the serial model, BACO is embedded in HRO as an operator to update the maintainer line. The experimental results on 0-1 KPs of different scales and correlations demonstrate the outperformance of the parallel model and serial model.
binary ant colony optimization algorithm is easy to be trapped into the local optimization region. The algorithm with controllable search bias is designed which fit for the unconstrained binary quadratic problem (UBQP...
详细信息
binary ant colony optimization algorithm is easy to be trapped into the local optimization region. The algorithm with controllable search bias is designed which fit for the unconstrained binary quadratic problem (UBQP). The complexity of the algorithm is analyzed. It is applied in UBQP, and the test set of ORLib shows the performance of the algorithm.
暂无评论