| Electromagnetism-like Mechanism (EM) Algorithm is a brand-new intelligentoptimization algorithm which has fast calculation speed and excellent global searchcapability. Since the algorithm was proposed, it has caused widespread concern byscholars at home and abroad. The algorithm has the advantages as fewer resourcesrequired, simple optimization mechanism, high convergence speed and strong searchability. However, EM is still easy to fall into a local optimal value, slow in convergencerate and bad in accuracy for solving some complex problems.The optimization principle of EM is studied deeply in this paper. An improvedElectromagnetism-like mechanism algorithm (IEM) is proposed here to eliminate theshortcomings of the original algorithm, such as bad quality of initial population, lowsearch efficiency, poor adaptability of step moving. IEM constructs the initialpopulation using disturbance factor and opposite learning mechanism not only to ensurethat the initial population contains rich patterns, but also to be conducive to theevolution of the overall convergence rate. It improves the local search algorithm withthe double chaotic search method that can increase the algorithm efficiency and avoidsinking into premature convergence; Considering the effects of iterations, an adaptiveformula of movement is designed, which brings improvement of accuracy and updatesthe particles move toward the optimal solution.Experimental results show that the improved algorithm has better optimizingperformance than the original algorithm. The new algorithm can improve not only theaccuracy of solutions, but also convergence efficiency, meanwhile the prematureconvergence is ameliorated.In further research, the EM will be studied further. The subsequent work will focuson improving the accuracy of the solution, and the application of the algorithm on morepractical engineering optimization problems, such as combinatorial optimization，recourses network and production scheduling. |
PDF Full Text Request