Research On CRO Algorithm For Solving TSP

In daily life, engineering design, facing many optimization problems,optimization problems is a research hotspot in computer science. When the scale ofthe problem is small, it is easy to get an accurate solution of the problem, but whenthe scale of the problem is large, you can use artificial intelligence methods to solvethe optimal solution, chemical reaction optimization algorithm is one of the intelligentalgorithms that have better performance and proposed in recent years.This paper use the TSP as the research object, CRO as intelligent algorithmframework for solve TSP problem, experimented four molecular manipulationoperators to adapt for TSP. This paper systematically summarized CRO algorithm anddescribed the molecular model, the molecular properties, the basic operation of themolecule, the basic idea and algorithm flow of CRO algorithms in detail. Based CROalgorithm process, this paper first initialized molecule collection, and for theshortcomings of first time initialization molecular uncertainty sampling, used arandom direction method to initialize a collection of molecules, me et the experimentsClaim of generation of molecules set with more information. Second, determined theCRO algorithm parameter configuration, affirmed the CRO algorithm functions. Forthe On-Wall Ineffective Collision in CRO, provide an optimized step for the On-WallIneffective Collision, to make up the defects of inadequate response of the originalOn-Wall Ineffective Collision convergence. For the decomposition reaction in CRO,provided a decomposition reaction step that can hold structure of the originalmolecule. For the Inter-Molecular Ineffective Collision in CRO, provided a newInter-Molecular Ineffective Collision step that can inherited structure of the originalmolecule, is easier to obtain the optimal solution than former. For the synthesisreaction in CRO, proposed a new synthesis reaction step that based on rotation, thisstep is more inclined to the local search.Through two experiments, proved CRO algorithm compare to the classical GAalgorithm is excellence at solve the symmetry TSP. CRO algorithms that optimizedfour reaction steps, can provide faster acceleration ratio for solve TSP. The largerscales the TSP have, the higher excellence the CRO algorithm have.