Cambridge scholar Xin-She Yang was inspired by the mechanism of flower pollination in nature,and proposed a new meta-heuristic algorithm—the flower pollination algorithm.The flower pollination algorithm includes two important search mechanisms,which are designed based on the flower pollination behavior of the flowering plant.Among them,a global search process is designed by simulating the flower cross pollination behavior,and a local search process is designed by simulating the flower self pollination behavior.The switching probability is designed to balance the strength between global search and local search,so that the algorithm can continuously iteratively search for the optimal solution of the optimization problem.Because of its simple concept,easy implementation,strong robustness,and high optimization rate,this algorithm has attracted great attention from many Chinese and foreign scholars after being proposed.At present,many scholars have improved the basic flower pollination algorithm by improving the parameters and incorporating the ideas of other evolutionary algorithms,improving and enriching the overall framework and search mechanism of the basic flower pollination algorithm.The optimized algorithm has different accuracy and convergence speed.To a greater extent,and can be effectively applied to a variety of complex optimization problems.However,with the deepening of the research,the researchers found that the algorithm has certain shortcomings,such as the lack of a reasonable design basis for the selection of parameter values in the algorithm,the algorithm's iteration to the later stage of the optimization performance weakened,unable to jump out of the local extreme value,and make flower pollination.It is difficult to expand the engineering application range of the algorithm.In this thesis,by studying and analyzing the basic flower pollination algorithm,we know that the performance of the algorithm mainly depends on the three aspects of cut flower probability,pollination method and evolution strategy.By improving these three aspects,the convergence speed of the algorithm is accelerated,the optimization precision of the algorithm is improved,and the overall optimization performance of the algorithm is improved.The improved algorithm is applied to the economic load allocation problem of power system to improve the optimization accuracy of the problem.The specific research work in this thesis mainly includes the following aspects:(1)An improved pollination algorithm based on Logistic chaotic mapping is proposed.The basic flower pollination algorithm uses a random method to initialize the population.This method makes the individuals of the population randomly distributed throughout the solution space,resulting in the algorithm's inefficiency in the optimization process.Chaotic mapping is sensitive to initial values,and has the characteristics of irregular order and ergodicity.Use these characteristics of chaotic mapping to initialize the population of the basic flower pollination algorithm,so that the individuals of the population are evenly distributed in the solution space,and can improve the exploration ability of the algorithm in the search space;the cross operator is introduced in the local pollination of the algorithm,Improve the algorithm's optimization accuracy,speed up the convergence speed,and appropriately enrich the diversity of the population.(2)An improved flower pollination algorithm based on the idea of differential mutation with dynamic switching probability is proposed.A dynamic switching probability is proposed to better balance the relationship between local search and global search,which promotes the overall performance of the algorithm.In the global pollination method of basic flower pollination,a differential evolution idea is introduced,a mutation operation is introduced,and two sets of random individual difference vectors are added.Maintain the diversity of the population,improve the global search ability,and improve the overall performance of the algorithm;introduce the target mutation strategy and set the local switching probability during the local pollination process,and use the local switching probability to achieve the target mutation strategy and the traditional differential evolution algorithm mutation strategy.The selection allows the algorithm to improve the local search convergence speed and maintain the diversity of local individuals in the population.(3)The optimization goal of power system economic load distribution refers to the rational distribution of load among power generating units,so that the target system can minimize the power generation cost under the conditions of meeting load requirements and operating constraints.The improved flower pollination algorithms are applied to the optimal allocation of economic load in the power system. |