关键词: 钢铁生产, 一体化批量计划, 粒子群优化, 约束满足, 邻域搜索

Abstract: The integrated batch planning of steel production continuous-casting hot-rolling was considered. By analyzing the essential operation properties of highlight problem, an integrated planning strategy was presented and a constraint satisfaction model was constructed. Considering the problem's NP-hard feature, a hybrid algorithm combining Improved Discrete Particle Swarm Optimization (IDPSO), constraint satisfaction and neighborhood search was proposed to solve this problem. With natural-number-matrix representation, each particle represented one solution to corresponding task allocation problem. Neighborhood search method was used to calculate particle's fitness value based on heuristic solution of the sort optimization problem. To improve the algorithm's convergence, the constraint satisfaction technique was employed to generate the initial feasible particle swarms and to revise unfeasible solutions during iterations. Validity of the model and algorithm were tested by calculating the data from production practices.

Key words: steel production, integrated batch planning, particle swarm optimization, constraint satisfaction, neighborhood search