Multi-objective Bee Clony Optimization Research And Its Application In Rolling Schedule

Artificial bee colony algorithm is a swarm intelligence algorithm based on bionics.Artificial bee colony has been applied to many fields because of its characteristics,such as simple structure,few parameters and strong global exploration ability,and so on.However,artificial bee colony algorithm still performs poor,such as easily to falling into local minima,weak exploitation,slow convergence speed,which have restricted its performance and applications.Therefore,the research on artificial bee colony algorithm has theoretical significance and practical value.On the basis of the deep study on artificial bee colony algorithm,two improved algorithms are proposed to solve the shortcomings.Moreover,the proposed algorithms are applied to load distribution for tandem cold rolling.The main works are as follows:(1)Considering the shortcomings of multi-objective artificial bee colony algorithm with strong exploration and weak exploitation ability,a multi-objective artificial bee colony algorithm based on limit search is proposed.The search radius is selected within the neighborhood according to the exploration of different honey,which balances the exploration and exploration ability of the algorithm,and accelerating the speed of convergence.Additionally,elite individuals in archive are given a fixed value,and the elite individual is selected according to the distribution and exploration of solutions in external archive.Based on the simulation of ZDT and DTLZ test problems,the approximate set of the Pareto front deduced by the proposed algorithm has better convergence and distribution.(2)Aiming at the problem of slow convergence of bee colony algorithm in processing complex multi-objective problems,a multi-objective bee colony algorithm based on regulation operator is proposed.The search radius is adaptively selected according to exploitation ability of honey,and focused on different search directions different periods.Besides,the fitness values of individuals are calculated in terms of distribution.Then the diversity solutions are selected by roulette wheel selection strategy to guide evolution.To verify the validity of the algorithm,UF test functions with complicated linkage variable are used for simulation,the algorithm performs well in convergence speed.(3)Based on the research background of five-stand cold tandem mill,by analyzing the model with the goal of preventing of skidding,equivalent rate of power,and good shape of final stand is established,and proposed algorithm are used to optimize the rolling schedule of tandem cold rolling.The simulation results show that the optimized load distribution scheme reduces the slippage problem,balances the load distribution,and improve the quality of the frame.