Microchip Manufacturing Plant Production Floor Scheduling

微芯片制造工厂生产车间调度

• 批准号: 2142174
• 金额: --
• 依托单位: Brunel University London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2018
• 资助国家: 英国
• 起止时间: 2018 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2142174
• 关键词: Microchip; Manufacturing; Plant; Production; Floor

Microchip manufacturing is a complex process that utilizes expensive machinery. To run process at maximum efficiency tight manufacturing schedules are used, to have minimum machinery down time and always have enough stock of demanded product. Occasionally microchip demand changes from existing or expected demand, and microchip production schedule has to be altered accordingly.The aim of this research is to further investigate Ant Colony Optimization algorithm's feasibility and effectiveness of solving Dynamic Multi Objective Optimization Problems by solving real world Dynamic Microchip Manufacturing Plant Production Floor Scheduling Problem. Investigate algorithm capability handle multiple aspects of dynamism, and varying degrees of dynamism. There has been no research done on solving optimization problems with more than one dynamic aspect of the problem yet.Develop and investigate performance of multiple multi objective strategies for given problemPareto optimal frontierObjective superposition Objective tier based compromiseDynamic optimization Robust dynamic optimization approaches are in particular interest for real-world applications where problems at hand have a tendency to evolve over time and are not predictable accurately in advance. Often such problems must have some mechanisms to improve solutions in real time as new data is obtained about disturbances or incremental accuracy improvements of the optimization data.Multi-objective optimizationIn real world optimization problems often have several conflicting objectives to be considered. Multi-objective results can be expressed in series of solutions that each one of them cannot improve any individual objective without reducing other objectives scores, this is called pareto optimal frontier. However, pareto optimal frontier is computationally expensive to build and rarely useful to have all pareto optimal solutions.Ant colony optimizationACO is metaheuristic optimization algorithm that mimics ant behavior of searching shortest path connecting their colony to source of food. Ants visiting route to the source of food lays down pheromone that attracts more ants to visit this path. Once the food source is depleted and disperse and pheromone evaporates therefore no longer attracting ants to unattractive path.Previous research has solved optimization problem for the pareto optimal frontier where decision maker has to decide which of the pareto optimal solutions to be used as a final solution, however knowing the domain of a problem, it is possible to enable algorithm find single best solution for a problem that can be automated and deployed in real time.

微芯片制造是一个复杂的过程，需要使用昂贵的机器。为了以最高的效率运行过程，使用了紧凑的制造时间表，以最小的机器停机时间，并始终有足够的所需产品库存。本研究的目的是通过求解真实的动态微芯片制造车间生产调度问题，进一步探讨蚁群算法求解动态多目标优化问题的可行性和有效性。研究算法处理动态性的多个方面和不同程度的动态性的能力。目前还没有研究解决具有多个动态方面的优化问题。针对给定问题开发和研究多个多目标策略的性能Pareto最优前沿目标叠加基于目标层的折衷动态优化鲁棒动态优化方法对真实的问题特别感兴趣。世界上的应用程序中，手头的问题往往会随着时间的推移而演变，并且无法提前准确预测。通常这样的问题必须有一些机制，以改善解决方案在真实的时间作为新的数据获得有关的干扰或增量精度的改进的优化data.Multi-objective optimizationIn真实的世界优化问题往往有几个相互冲突的目标要考虑。多目标结果可以表示为一系列解，其中每一个解都不能在不降低其他目标得分的情况下提高任何单个目标，这称为帕累托最优边界。蚁群优化算法是一种模拟蚂蚁搜索连接其群体和食物源的最短路径的元启发式优化算法。蚂蚁到食物源的访问路线会产生信息素，吸引更多的蚂蚁访问这条路线。一旦食物源耗尽和分散，信息素蒸发，因此不再吸引蚂蚁到没有吸引力的路径。以前的研究解决了帕累托最优边界的优化问题，决策者必须决定哪一个帕累托最优解作为最终解，然而，知道问题的域，使得算法能够找到问题的单个最佳解决方案是可能的，该解决方案可以自动化并真实的部署。