An Improved Continuous Tabu Search Algorithm with Adaptive Neighborhood Radius and Increasing Search Iteration Times Strategies

Abstract

Based on Continuous Tabu Search with Sine map and Staged strategy (CTS-SS) algorithm, an improved CTS-SS (iCTS-SS) algorithm is proposed to achieve rapid convergence and enhance the robustness of the CTS-SS algorithm with strategies of adaptive neighborhood radius and awarding search iterations. Unlike a fixed neighborhood radius used in the CTS-SS algorithm, the iCTS-SS algorithm utilizes the adaptive neighborhood radius to improve the search capabilities. According to whether the optimal result is replaced, it could be narrowed for centralized search or expanded for decentralized search. By this way, the result can be better than using regular iterations. The experiments adopt various typical functions as numerical experiment cases to evaluate the robustness, accuracy and efficiency of iCTS-SS. And for comparison, these functions are also solved by CTS-SS and Modified Staged Continuous Tabu Search (MSCTS), which is one of the latest continuous tabu search algorithms. Experimental results show that iCTS-SS outperforms both CTS-SS and MSCTS in statistical significance.