Optimization is a widely used subject. Electromagnetism-like mechanism (EM) method was brought forward by Birbil and Fang in 2002, being inspired by eject-attract mechanism of charged particles in electromagnetic field. It is a new population-based heuristic algorithm. Its idea is to simulate the eject-attract mechanism of charged particles in electromagnetic field and make every particle move to the optimal solution by established some determinate rules, and then obtain the optimal solution of the problem. In this paper, electromagnetism-like mechanism method is introduced in detail. Basing on the fundamental framework of the Electromagnetism-like mechanism algorithm, the local search method of the algorithm is improved. Formula of resultant force is redefined and a mutation operator is introduced to accelerate the convergence speed of the algorithm. The results of the numerical experiments indicate that these improvements accelerate the speed and increase the efficiency of the improved algorithm. On the other hand, while penalty function is designed after using the maximum entropy method simplify the constraint conditions, the constrained optimization problems are transformed into non-constrained optimization problems. Satisfied results are obtained by using the improved electromagnetism-like mechanism method to solve some constrained optimization benchmark problems, it showed that constrained optimization problems can be solved by the improved algorithm. |