Abstract: In a continuous production environment, in order to minimize the make-span, the vehicle sharing between raw material procurement and finished product delivery is considered, the joint scheduling model of raw material procurement-continuous production-finished product distribution is constructed and the correspondence between problems and algorithms is analyzed, an improved artificial bee colony algorithm is designed. In the algorithm, firstly two-dimensional matrix encoding is used to represent the combination of vehicle and material priority value. To differentiate the matrix initial value, a transfer function that transforms random number between（0,1） to the initial value between[(0,+∞)]is desinged. In the decoding process, phased decoding is used to ensure the continuity of production. Secondly, in order to realize the efficient use of vehicles, heuristic information for vehicle selection is designed. Finally, through the test cases and comparative experiments, it shows that the effectiveness of the joint scheduling strategy is verified compared to the traditional scheduling strategy. When the material demand is increased, compared with the overall decoding method, the stepwise decoding method can obtain a feasible solution. The heuristic information enables the improved artificial bee colony algorithm to improve the effect of the solution by 0.63%, compared with the genetic algorithm, the improved artificial bee colony algorithm has an average improvement of 0.35%.

Key words: continuous production, joint scheduling, car sharing, bee colony algorithm

摘要： 在连续生产环境下，以最小化完工时间为目标，考虑了原料采购与成品配送过程的车辆共用问题，建立了原料采购-连续生产-成品配送的联合调度模型，并分析了问题特性，设计了改进的人工蜂群优化算法。采用二维矩阵的编码方式表示车辆与物料组合优先权值。为使矩阵初值差异化，设计了将（0，1）间随机数转化成[(0,+∞)]实数值初值的转换函数。在解码过程中，为保证生产的连续性，采用分阶段解码的方式。为实现车辆的高效利用，设计了车辆选择的启发式信息。通过测例及比较实验表明：相对于传统的调度策略，验证了联合调度策略的有效性；在增大物料需求量时，相对于整体解码方式，分步解码方式能得到可行解；启发式信息使改进的人工蜂群算法求解效果提升了0.63%，改进的人工蜂群算法相对于遗传算法，求解效果平均提升了0.35%。

关键词: 连续生产, 联合调度, 车辆共用, 人工蜂群算法