Research On The Batch Scheduling Method Based On ?/?-Sequence Rule For Mixed-Model Assembly Line

In order to adapt to the change of production mode from mass production to the personalization customization, more and more manufacturing enterprises turn to organize production according to orders from customers, thus there's a very need to produce different products in a same line. Complex equipment always has a complicated product structure due to the variety of components needed for assembly, the BOM of different series of products or different models in a single series has both differences and similarities with each other. Thus workshop usually set products with same or similar machining characteristic or product structure into production batches when making a production plan. Under batch production mode, on-time order delivery is hard to be guaranteed. Meanwhile, stagnation of assembly line often appears due to interference factors such as incomplete components supply, consequently deteriorating the production efficiency and resulting in excessive overtime. Thus the optimal scheduling of a single assembly line is the basis to ensure the smooth production of the mixed-model assembly system of complex equipment, as well as the premise to improve the quick response ability to the change of demands and to realize the lean controls of manufacturing enterprises.This paper considers the problem of sequencing items for mixed-model assembly, which assembles various models in a single line and batches in shifts. Firstly this paper expounds the production characteristic of no-wait mixed-model assembly line, summarizes the research status of scheduling models with batching constraints and analyzes the problems in the practical production of mixed-model assembly make-to-order mode. And then a multi-objective sequence model considering above factors is established with the objectives of minimizing processing cycle, total earliness/tardiness time and overtime. A "?" Takt-based sequencing rule is proposed to obtain the better outcomes of sequencing. Finally, since the sequencing problem typically provides an intractable search space for problems of'real world' size, a weight-space ant colony (WSACO) algorithm is designed and tested its performance on randomly generated instances. Besides, experimentation shows that the ?-sequence rule, and ?-sequence rule (the extension of ?-sequence rule) outperforms SPT and LPT in shortening processing cycle and provides a new method for solving multi-objective batch scheduling problems.