Application Of Improved Harmony Search Algorithm In Multi-objective Optimization

In the social life and engineering practices,there are often many optimization problems that can be generally divided into two categories,namely single-objective problems and multi-objective problems,the later of which is more common in living practices.With the improvement of people's requirements on social life and the engineering quality in industrial field,it is more and more complicated for multi-objective optimization problem,and the multi-objective optimization algorithms for solving multi-objective problems are also getting more and more attentions as well as are studied by experts and scholars.Aiming at the multi-objective optimization problems,the thesis carries out research and exploration with Harmony Search Algorithm(HS).The Harmony Search Algorithm(HS)has the advantages of few algorithm parameters,simple implementation,and easy integration with other intelligent algorithms.At meanwhile,there are also many disadvantages,such as it is easy for the algorithm to fall into the local optimal,and it has high requirements on the initial solution set diversity.The thesis makes improvement on the shortcomings of Harmony Search Algorithm(HS)and applies it into the multiobjective optimization problems in real life.The research contents are as follows:At first,the thesis introduces the standard Harmony Search Algorithm(HS)and multiobjective optimization problems.It also researches and concludes the characteristics,advantages and disadvantages of the algorithm,as well as its actual engineering application fields.At the same time,the thesis also introduces the mathematical form and solving process of multi-objective optimization problems,as well as the optimal related concepts of Pareto and several classic multi-objective optimization algorithms.Then,aiming at the disadvantages of the algorithm,the thesis puts forward three improvement points for the algorithm through combining with its characteristics: 1.aiming at the problem that it will have a influence for the solution site diversity in the initial harmony memory library of the standard Harmony Search Algorithm(HS)on the optimization process of the algorithm,the thesis introduces the Tent chaotic system to generate the initial solution set and strengthen the distributing uniformity of the initial solution set in the range of decision variables,in order to improve the optimization process of the algorithm.2.Aiming at the problem that it is sensitive for the standard Harmony Search Algorithm(HS)to the parameters,the thesis dynamically adjusts the algorithm parameters,in order to make the main parameters of the algorithm dynamically adjust with the optimization process of the algorithm.In the early stage,it strengthens the global searching ability of the algorithm,and in the later stage,it emphatically improves theability of algorithm for jumping out of the local optimal,in order to achieve the comprehensive improve on the characteristics of the algorithm.3.For multi-objective optimization problems,the thesis improves the elite selection strategy in the algorithm through combining the characteristics of the algorithm with easy integration with other intelligent algorithms as well as integrating the fast non-dominated sorting genetic algorithm(NSGA-II),which adopts a circular elimination method to strengthen the uniformity of the first-layer non-dominated solutions and improve the Pareto optimization ability of the algorithm.At last,the thesis carries out the numerical simulation experiments on the improved algorithm and other commonly used multi-objective optimization algorithms through the standard test function.With the comprehensively comparative analysis,the result shows that the improved algorithm in the thesis can effectively avoid falling into the local optimum,but with a better optimization accuracy.The thesis applies the improved Harmony Search Algorithm(HS)into the vehicle routing problem of multi-objective cold-chain logistics,and then conducts a mathematical modeling from the aspects of total distribution cost and customers' satisfaction.In addition,it also adjusts and applies the improved algorithm according to the characteristics of the problem itself,in order to analyze the solving results of the algorithm,which can be seen that it is effective and feasible for the improved Harmony Search Algorithm(HS).The last is the conclusion in the thesis,which makes clear the further research directions.