The Modified Whale Optimization Algorithm And Its Applications On Function Optimization And Notwork-on-Chip

With the increase of complexity of optimization problems,the scale of the optimization problems is exploding as a result of the more unknown variables,the complicated structures of objective functions,the huge constraints,the high dimension and the complex relationship between the variables.Therefore,how to address the complex optimization problems is becoming more and more difficult and challenging.As a kind of heuristic algorithms,swarm intelligence algorithms have been widely applied to various kinds of optimization problems and achieved success because of its simple parameters,not depend on the specific optimization problem,strong global search ability and learning ability.Whale optimization algorithm is a new group-based metaheuristic algorithm which has been proved to be superior to other group-based intelligent algorithms in some optimization problems and engineering applications.This paper studies whale optimization algorithm and applies the improved whale optimization algorithm to high-dimensional function optimization problems and network-on-chip(NoC)mapping optimization problems.The major job and contribution described as following:(1)A briefly summary and analysis of the optimization problem are given,in which high-dimensional function optimization problems and network-on-chip mapping optimization problems are introduced respectively.The former is a kind of continuous optimization problem and the latter is a kind of discrete optimization problem.The background and development status of the two optimization problems are described in detail,and the existing results are analyzed and evaluated.In addition,according to different large-scale optimization problems,the problems and challenges to be solved are pointed out.(2)The classical swarm intelligence algorithms are elaborated and compared,and the biological background and bionics principles are given.The whale optimization algorithm is analyzed in detail,and the characteristics and functions of different search strategies in the process of optimization are summarized.At the same time,its limitations in large-scale optimization are analyzed.In order to address the limitation of whale optimization algorithm,the improvement strategies in term of exploration and exploitation are introduced.(3)This paper proposes an improved whale optimization algorithm(MWOA)and proves its effectiveness by solving large-scale function optimization problems.The proposed algorithm adopts quadratic interpolation operator to increase the diversity of the population and guides the search agent to conduct fine search in the neighborhood of the current optimal solution so that can help the algoriths enhance the exploration ability of the algorithm.Using Levy's flying short distance walk and occasional long distance jump to jump out of the local optimal algorithm to avoid premature convergence;The search process is controlled by using nonlinear parameters instead of linear parameters to accelerate the convergence of the population and improve the balance between exploration and exploitation.Simulation results show that compared with the four new improved algorithms,the proposed algorithm is very effective for solving high-dimensional function optimization problems in terms of convergence speed,optimization accuracy and overall performance.(4)The mapping problem of network-on-chip requires one-to-one correspondence between IP cores and network nodes,so this optimization problem is discrete and has constraints.In order to solve large-scale NoC mapping optimization problems effictively,a mapping method based on genetic algorithm and whale optimization algorithm(WOAGA)is proposed,which has a good stability and can minimize energy consumption.The improved two-point crossover operator is used to exchange information and increase the diversity of the population,while the mutation operator is used to introduce new genes to avoid the search agent stagnating in the local optimum.Combined with the parameter design of the whale algorithm,the algorithm changes the focus of the search along with the search process,so as to better meet the requirements of different search stages.Simulation results show that the proposed algorithm is effective in solving the optimization problem of large-scale NoC mapping problems.