一种分片混沌贪婪振荡退火 TSP 优化算法.

The simulated annealing algorithm with adaptive temperature rising strategy or Monte Carlo strategy has the problems of slow convergence and limited global optimal approximation ability in solving complex 'ISP respectively. On the other hand, the existing chaotic optimization algorithm has the defect of logistic mapping, which weakens its ability to jump out of local optima. Therefore, this paper designed a fusion algorithm framework, in which embedded the divided Lorenz chaotic mapping system to enhance the search efficiency of the chaotic algorithm for the neighborhood solution. It introduced the greedy strategy to construct the initial solution approaching the global optimal solution, which made the algorithm had the ability of transition to the neighborhood of the global optimal solution. In addition, it designed the complementary mechanism of oscillatory annealing to improve the sub iterative solution screening process, and enhanced the global search performance of the algorithm. After the implementation of the algorithm, using the international public 1SPLIB point set, through multiple rounds of comparative testing, it verifies that the performance of the new algorithm is better than the comparison group simulated annealing algorithm and logistic chaos optimization algorithm,and has shorter convergence time and stronger global optimal approximation ability. [ABSTRACT FROM AUTHOR]