Study On Rolling Load Distribution Based On Intelligent Optimization Method

The setting of rolling load distribution is important to strip steel.In the past,the load distribution methods based on experience have been replaced by intelligent optimization algorithms.Based on Glowworm Swarm Optimization(GSO),the load distribution is set and its shortcomings are improved to deal with the load distribution problem better.To solve the problems of slow convergence and easy to fall into local extremum,an Immune Glowworm Swarm Optimization(IMGSO)is proposed.It uses the feature of Immune algorithm and new fluorescein formula.The results show that convergence times of IMGSO are 17 times less than those of GSO,the time is reduced by 0.98 s,there is no redundant iteration in the convergence curve.To solve the problems of slow convergence and poor local optimization ability,Variable Step Glowworm Swarm Optimization(VSGSO)is proposed.It reduces the step size by information of the distance between fireflies.The results show that the convergence times of VSGSO are 9 times less than those of GSO,the time is reduced by0.52 s,the convergence accuracy and variance are both improved.To solve the above problems,Immune Variable Step Glowworm Swarm Optimization(IMVSGSO)is proposed,which combines IMGSO with VSGSO.The results show that the convergence times of IMVSGSO are 2 times less than those of IMGSO,the time is reduced by 0.4s,the convergence accuracy and variance are equivalent to VSGSO,there is no redundant iteration in the convergence curve.To solve the problem of poor Pareto frontier,Chaotic Glowworm Swarm Optimization(CGSO)is proposed.The algorithm uses Logistic chaotic map to weight multi-objective function.The results of ZDT sequence show that the Spcing and Generation Distance of CGSO algorithm are smaller than NSGAII;The results of load distribution shows that solution of CGSO dominates the solution of NSGAII,which verifies the superiority of the proposed method.Figure 19;Table 13;Reference 53...