基于变量分离的生产调度空间划分算法

静态生产调度大多形成MILP或MINLP模型，由于调度规模大及混合整数规划的组合优化特性，造成调度求解困难。通过对混合整数规划模型空间的分析，提出依据整数变量和连续变量的分离策略进行空间的自然划分，从而将模型的求解转化为多个较小规模连续子空间的寻优。对典型间歇调度模型的分析表明，将空间划分后进行连续寻优的策略较大地降低了实际运算的规模，降低了求解难度，可以提高问题的求解速度和效率。

Variable-separating-based Space Decomposition Algorithm of Production Scheduling

Most static production scheduling problems are formulated in MILP(mixed integer linear programming) or MINLP(mixed integer non-linear programming).It is difficult to find solutions of scheduling because of its large scale and combinatorial characters.According to the features of MIP(mixed integer programming),integral variables and continuous variables are separated and the searching space is decomposed naturally into many continuous subspaces of less scale.Taking a typical batch production scheduling as a case,the analysis shows that variable-separating strategy decreases the scale of continuous searching problem greatly and makes it easy to solve,which can improve the speed and efficiency of optimization.