Abstract: Due to the isolation between nodes in the set of nodes and low searching efficiency of local short-range connection in the simple small world algorithm, which lead to trapping into local minimum when dealing with complex functions, a?new small world algorithm with balanced?optimization strategy is proposed. The number of candidate nodes, searching probability of local short-range connection of the originating nodes and the hamming distance between the originating and their candidate nodes is changed dynamically during optimization, while elite strategy and new searching strategy of local short-range connection is used, in order to insure reasonable balance of exploration and development for the search space in different optimized period. Simulation results show that convergence, reliability, efficiency and stability of the new algorithm are superior to the original. The new one effectively solves the stagnation and low searching efficiency of local short-range connection, being endowed with the potential to solve complex engineering optimization problems.

Key words: small world algorithm, optimization?strategies, reliability, optimize efficiency

摘要： 针对简单小世界算法由于节点集中节点之间信息孤立，以及局域短连接搜索效率较低的原因，造成在优化复杂函数时出现的停滞现象，提出了一种基于均衡性策略的小世界优化算法。在优化过程中，动态改变源节点的候选节点数、短连接搜索概率和与候选节点之间的海明距离，使用精英策略和新的短连接搜索策略，来保证在不同优化时期，对探索空间的“探索”作用和“开发”作用的合理权衡。仿真试验表明，在收敛精度、收敛可靠性、收敛速度和收敛稳定性等方面都优于原算法，有效地解决了原算法的停滞现象和低局域短连接搜索效率的问题，具备解决更复杂工程优化问题的潜能。

关键词: 小世界算法, 优化策略, 可靠性, 优化效率