Optimization Of Multi-Objective Rolling Schedules For Hot Rolling Aluminum Mill Based On Improved Particle Swarm Optimization Algorithm

This paper took the “1+4” aluminium strip finishing rolling mill of a factory in Henan as the research object, and had a depth research for the design of the rolling schedule. The design of the rolling schedule is the core content of aluminium strip production process, and multiple objective functions should be considered in the process of making a rolling schedule. Traditional single target rolling schedule design method is difficult to meet the site requirements, so the rolling schedule of the multi-objective optimization problem to be solved.Firstly, in view of shortcomings that can’t guarantee global optimization and a good diversity of particle swarm optimization algorithm in solving multi-objective problem, a multi-objective adaptive chaotic particle swarm optimization(MACPSO) is proposed. The algorithm adopted a new method to select the global optimum particle, improved external archive updating strategy, and an adaptive mutation strategy was presented for the external archive. The experimental results show that under the premise of guarantee convergence speed, the proposed algorithm has a better convergence and distribution, showing excellent optimization performance.Then, a depth analysis of aluminum hot rolling mechanism model is carried on and the deformation resistance regression model is regressed by the data from the field. On this basis, the multi-objective function model with equal power margin and preventing the slip phenomenon is established and it’s optimized by MACPSO. A high quality set of rolling schedule is got, and have a comparative analysis with traditional methods.Finally, an selection strategy based on the fuzzy preference is proposed for the problem of the finally rolling schedule choice. In the method, the fuzzy preference for the objective function of decision-makers is ranked and mapped to a specific rolling schedule in the solution set, there is provided a decision-making method for the level two system.