间歇蒸煮过程的分层多目标优化

Lexicographically Stratified Programming for Batch Cooking Process

对制浆生产间歇蒸煮过程进行了详细分析 ,建立了包括环境污染为目标函数的多目标分层规划数学模型 ;将传统分层多目标优化的方法和遗传算法相结合 ,提出了一种新的多目标优化方法。遗传算法中综合了并列选择和小生境方法 ,采用了浮点编码和算术交叉操作 ,应用罚函数法将有约束优化问题转化成无约束优化问题。应用实际间歇蒸煮过程数据的优化仿真结果表明 ,该方法可以在保证纸浆质量的前提下 ,减少排放废液中碱的含量 ,降低生产成本 ,达到清洁生产的目的。

The cooking process produces a large amount of waste and pollution, which become one of the main pollution sources in pulp and paper industry. The paper studied the model of a batch cooking process based on analyzing the process in details. A lexicographically stratified programming mode containing pollution objective function was established first. A new hybrid lexicographically stratified programming mechanism was proposed, it combined multi-objective with genetic algorithm. Parallelism selection and niche method were generated, and floating point representation and arithmetic crossover were used in the genetic algorithm, constraint domain was transformed to no constraint by using penalty function. The results of computer simulation based on the real data from a pulp mill shown that by using the method suggested, the pollutant can be reduced effectively and the profit will increased while keeping the quality of the pulp.