基于多目标免疫算法和数学规划的双行设备布局方法研究

Facility layout problems exist widely in the field of manufacturing and have important academic significance and realistic value. Currently,researches on single row and multiple rows facility layout problems，which have been applied widely to the area of manufacturing, only need to determine the sequence of machines and do not consider their accurate locations. Those researches usually use the material handling cost as the only optimization objective and seldom consider the objective of layout area. However, the construction costs of workshops for some industries are very expensive, such that the layout area is also an important factor to be considered; in addtion, considering the accurate locaitons of machines is also benefit to obtain a facility layout with a lower material handling cost. Hence, this project researches the double rows facility layout problem that considers the accurate loactions of machines and consists of two objectives of cost and area, and then proposes a new approach that integrates a multi-objective immune algorithm with the mathematical programming to solve this problem. This research includes: (1) establish the mathematical programming formulation of this proposed double rows facility layout problem; (2) research the mathematical programming based method to costruct the non-dominated set of solutions for an arbitrary sequence of machines; (3) research the multi-objective immune algorithm based method to obtain the non-dominated set of machines sequences; (4) Research the integration of the multi-objective immune algorithm and the mathematical programming approach to achieve the Pareto optimal set of solutions; (5) further, propose a double rows facility layout problem with additional clearances, and then research the approach to solve this problem.

设备布局问题在制造领域中广泛存在且具有重要理论意义与应用价值。当前制造领域广泛应用的单行或多行设备布局研究中，仅需确定机器的排序，而不需考虑机器的精确位置，优化目标大多为材料处理成本，而很少把布局面积作为优化目标。然而，某些制造领域的车间建造成本非常高昂，因而布局面积也是需要考虑的重要因素，并且考虑机器的精确位置能获得成本更低的设备布局。为此，本项目研究考虑机器精确位置的带有成本和面积两个优化目标的双行设备布局问题，提出一种结合多目标免疫算法和数学规划的新方法来解决该问题，具体包括：（1）建立所提出的双行设备布局问题的数学规划模型；（2）研究基于数学规划来构造任一机器排序的非支配解集的方法；（3）研究基于多目标免疫算法来获取非支配的机器排序集合的方法；（4）研究结合多目标免疫算法与数学规划方法来获取Pareto最优解集的方法；（5）进一步提出带有附加间距的双行设备布局问题并研究其解决方法。

设备布局问题在很多领域中广泛存在，对于提高生产效率、节约成本具有重要意义。本项目主要研究考虑设备精确位置的带有成本和面积两个优化目标的设备布局问题，内容包括：（1）研究了考虑设备精确位置的双行设备布局问题。建立其数学规划模型，提出任一设备排序的非支配解集的构造方法，进而提出一种结合多目标优化算法（禁忌搜索和免疫算法）和线性规划的新方法来解决该问题。（2）研究了带有共享间距的双行设备布局问题。通过考虑相邻设备间的共享间距来获取更紧凑的设备布局，以减少物料流成本。建立其数学规划模型，提出一种结合多目标禁忌搜索和数学规划的求解方法。（3）研究了一种半导体制造领域中的双贝布局问题。该问题需确定设备的排序和精确位置，利用单向导轨在设备间传送材料。建立了问题的规划模型，提出一种结合多目标遗传算法和线性规划的方法来同时优化材料流成本和布局面积。（4）研究了动态(鲁棒)双行设备布局问题。该问题包含多个生产时间段，设备间物料流在各时间段发生变化。建立其数学规划模型，提出基于模拟退火算法和多目标进化算法的求解方法。（5）研究了医院科室布局问题。以病人行走距离和科室邻近需求作为优化目标，提出一种基于多目标进化算法的医院科室布局方法。经该方法优化的科室布局能显著减少病人行走时间和距离。