The Improvement And Application Of Chemical Reaction Optimization Algorithm

Optimization problems are prevalent in scientific research and engineering.However,the traditional optimization algorithm cannot obtain satisfactory solution in the short time when it faces the optimization problem with higher complexity,stronger binding,and more nonlinear.The intelligent optimization algorithm based on the biological intelligence or physical phenomena and has the characteristics of random search shows superior performance when solving complex optimization problems,and has aroused widespread concern of scholars both at home and abroad and is applied to optimize optimization issues in various fields.However,the existing intelligent optimization algorithm does not solve all optimization problems well.Therefore,it has great academic and engineering value to study new or improve existing intelligent optimization algorithms.In this paper,we focus on chemical reaction optimization algorithm,and propose an improved real-coded chemical reaction optimization algorithm(ICRO)whose performance is verified by solving the global numerical optimization problem and two chemical reaction optimization algorithms based on different solutions(VMP_CRO_P and VMP_CRO_V)for solving the virtual machine placement problem in cloud computing.The main work is as follows:(1)Aiming at the problem that the real-coded chemical reaction optimization algorithm solves the problem of global numerical optimization,it has a lack of convergence of low precision,slow convergence.In this paper,an improved real-coded chemical reaction optimization algorithm(ICRO)is proposed,which incorporates the information exchange mechanism between molecules in the population and improves the synthesis operation of the original algorithm for intermolecular collision operation.The experimental results show that ICRO has better convergence accuracy,stability and faster convergence rate,and the improved part of the ICRO algorithm can be incorporated into other hybrid algorithms(such as HP CRO and OCRO)to further improve the overall performance of the hybrid algorithm.(2)Based on the superiority of chemical reaction optimization algorithm in combinatorial optimization problem,two chemical reaction optimization algorithms(VMP_CRO_P and VMP_CRO_V)are proposed to optimize virtual machine placement problem(VMP).Virtual machine placement problem is a hot research problem in the field of cloud computing.It is very important to reduce the energy consumption of data center,improve the utilization rate of physical resources and the sustainable development of data center.VMP_CRO_P and VMP_CRO_V have different solutions and different basic operation operators.Simulation experiments show that both algorithms can effectively reduce the energy consumption of the physical server and improve the resource utilization of the physical server.The algorithm VMP_CRO_V has better scalability.