Modeling And Optimization Study Of Hybrid Production Scheduling

With the rapid development of manufacturing, modern enterprise's product structure changes from single production to hybrid production. This production way contains continuous production and discretion production, which is belonging to Hybrid Dynamic Systems (HDS). The production scheduling of this system is NP-hard problem, modeling and solving process are very difficult.This paper mainly research about how to simply modeling process through staging method, and how to optimize scheduling process on the modeling. At the same time, discuss the dynamic scheduling problem in real produce process under the influence of emergent parts, machine repair and so on.The main study of this paper as follows:1. First of all, this paper collects and analyzes the research development of domestic and abroad hybrid production scheduling problem, giving a summary on the methods of production scheduling, focusing on the development and exist problem, provide theory method for establishing method model and optimize algorithm.2. This paper takes a class of hybrid production process with buffer for example, analyze its production process and do some assumptions about part modeling condition. And this production scheduling is separated two stages, establishes multi-objective optimization model of production costs minimum, the biggest advance/maximum tardiness penalties and minimizing the maximum completion time.3. Based on the scheduling rules, optimizes hybrid production scheduling problem using Heuristic branch and bound method, and gets the optimal production scheduling program. This algorithm not only contains Batch process for production tasks, but also discusses the sequencing problem of discrete production process.4. According to the characteristics of hybrid production and dynamic scheduling, this paper puts forward a multi-objective genetic optimization strategy based on the rolling window. This strategy based on periodic and event-driven rescheduling mechanism of hybrid multi-objective(Delivery time, maximum flow through time, the maximum completion time and deviation from the initial scheduling) evolutionary algorithm for scheduling optimization of window parts.5. Through calculating the example, inspect and verify the statistic and dynamic model and algorithm. The model result demonstrates that this model effective simply the complexity of hybrid process system. At the same time are practical and feasible.