Improvement And Application Of Sine Cosine Algorithm

Nowadays,the rapid development of social industry,economy,and technology has spawned a large number of high-dimensional,complex,and nonlinear optimization problems.The shortcomings of classical algorithms in solving high-difficulty optimization problems,such as high time complexity,poor calculation results and slow convergence speed,make swarm intelligence algorithms get an opportunity to develop.Swarm intelligent algorithm has been favored by many experts and scholars for its advantages of faster convergence speed,easier implementation and independent of the inherent characteristics of the problem.After years of research by scholars at home and abroad,swarm intelligent algorithms have been greatly developed,and the number of algorithms has been greatly expanded.Sine cosine algorithm(SCA),as one of the new swarm intelligence algorithm,its application and algorithm improvement related papers have been published continuously.However,most of its applications are focused on numerical optimization problems.There are few papers using SCA to solve combinatorial optimization and mixed integer optimization problems.Therefore,it has certain research value to use SCA to solve bounded knapsack problem(BKP),knapsack problem with a single continuous variable(KPC)and optimal reactive power dispatch(ORPD)of distribution network.In order to solve BKP with SCA,a discrete sine and cosine algorithm(DSCA)is proposed based on the code conversion method.In DSCA,individuals use integer coding methods to represent BKP solutions,and use repair and optimization method to eliminate the infeasible solutions of BKP generated by the algorithm.For the three large-scale BKP instances,comparison with the pigeon inspired optimization and particle swarm optimization shows that DSCA not only has better calculation results,but also has excellent stability,indicating that DSCA is a new effective algorithm for solving BKP.Based on the discrete KPC model and the two-subproblem KPC model,the four-subproblem KPC model KPCM4 is generalized.In order to verify the performance of SCA in solving KPC and the advantages and disadvantages of different models,single-population SCA uses discrete KPC model,double-population SCA uses KPCM3 model and multi-population SCA uses KPCM4 model to simulate four kinds of KPC instances.The results show that the KPCM4 model is slightly better than the KPCM2 model and discrete KPC model and SCA is a new method to solve KPC.To verify the performance of SCA in solving discontinuous numerical problems,SCA is applied to solving ORPD problems.The pseudo code description and flow chart of SCA for reactive power optimization are given,and the IEEE 30-bus instance is simulated.The network loss after optimization is compared with that before optimization.The experimental result shows that SCA can effectively optimize the voltage value of each node and reduce the active power loss of distribution network.It also shows that SCA can be used to solve mixed integer optimization problems.