Improvements And Applications Of Quantum Evolutionary Algorithm In Rolling Schedule Multi-objective Optimization

Rolling schedule is the provision of reduction and rolling force of each rolling racks inthe rolling process. The various factors need to be considered in rolling scheduleoptimization such as rolling energy, balance of equipment load, shape, gauge and so on. It isdifficult to obtain reasonable optimization results by traditional single-objectiveoptimization.. Multi-objective optimization algorithm has many advantages in solvingmulti-objective optimization problem. The improvement on multi-objective optimizationalgorithm to adapt to rolling schedule optimization is one of hotspot issues in domestic andforeign scholars.On the basis of rolling production process, modeling and design method for rollingschedule optimization model are discussed, and the empirical formula of rolling scheduleoptimization are provided such as optimization function, decision variables and constraints.Based on the theory of quantum computing and the concept of multi-objective evolutionaryalgorithm, the verified solution is selected as multi-objective quantum evolutionaryalgorithm. The treatment idea on rolling schedule multi-objective optimization problem ispresented.For the features of hot rolling schedule multi-objective optimization, multi-objectiveoptimization genetic algorithm based on Q-bits real-coded is proposed. The algorithm usesquantum-bits real coding to solve complicated problems in binary encoding and decoding,interferes crossover and mutation of genetic operator by the quantum state, considersnon-dominated solutions and selects populations based on crowding distance. And the globalPareto optimal solution set is ensured. Hot rolling schedule multi-objective optimizationmodel is established according to seven rolling machines in actual production process of hotrolling factory, and is optimized based on the proposed algorithm, and then better rollingschedules are obtained compared with original schedule, the convergence speed ofmulti-objective evolutionary algorithm in hot rolling schedule multi-objective optimizationis accelerated.For the features of cold rolling schedule multi-objective optimization, quantummulti-objective immune algorithm based on adaptive preference is proposed. Adaptivepreference learning mechanism is introduced to the proposed algorithm, so high quality Pareto solution set is achieved. Cold rolling schedule multi-objective optimization model isestablished according to five rolling machines in actual production process of cold rollingfactory, and is optimized based on the proposed algorithm, and then better rolling schedulesare obtained compared with original schedule, the convergence and the distribution ofmulti-objective evolutionary algorithm in cold rolling schedule multi-objective optimizationis improved.