The Study Of Artificial Bee Colony Algorithm Based On Dynamic Penalty Function And Multi-objective To Solve Constrained Optimization Problem

Artificial bee colony algorithm (ABC) is a recently proposed in-telligent optimization algorithm. Due to its advantages with simple,less control parameters, and easy to implement, it is widely applied in various fields of social science. The constrained optimization prob-lem widely exists in social life, using artificial bee colony algorithm to solve the complex constrained optimization problems is an impor-tant research topic. In all kinds of improved versions of artificial bee colony algorithms, there are still some shortcomings. In view of this,this paper puts forward two kinds of improved artificial bee colony algorithms for solving constrained optimization problems.Firstly, for the improved artificial bee colony algorithm (CABC),which is easy to fall into local optimum, this paper puts forward modi-fied artificial bee colony algorithm based on dynamic penalty function and Levy flight (DPLABC). The algorithm replaces the Deb's rules with dynamic penalty function method to prevent algorithm falling into local optimum, and introduces search mechanism based on Levy flight, which balances the exploitation and exploration well in two stages. In onlooker bee phase, DPLABC algorithm makes full use of the following probability, and puts forward blended strategy learning from the best solution and multiple neighborhoods. 24 benchmark functions and 4 classic engineering design problems are tested, and compared with the state-of-the-art algorithms. The results show that DPLABC algorithm has obvious advantages.Secondly, inspired by combining with multiobjective optimiza-tion differential evolution algorithm (CMODE), this paper proposes the modified artificial colony algorithm (MONABC) based on multi-objective and non-dominated solution replacement mechanism. In employ bee phase, the algorithm adopts multiple search mechanisms in the small population, all kinds of group has different search mecha-nism. Non-dominated solutions in the offspring population randomly replace an its dominated solution in original population, which up-dates population. Moreover, the following probability formula based on the sorting method is establishd, and on the scout bees stage, we will improve the replacement mechanism from CM ODE algorithm.MONABC algorithm is tested on 24 benchmark functions, the ex-perimental results show that the algorithm has certain competitive.