Search

In the construction industry, delivery times of building materials are often delayed to fit the progress of construction, and therefore parts manufacturers are forced to hold extra parts inventory over the period of the delivery delay. To minimise the inventories, two production models, i.e., push and pull type production models can be widely considered. When production lead-time is short and length of delivery delay is long, a pull type may require smaller inventories to keep a certain service level, and a push type may require smaller inventories under the condition of longer production lead-time and shorter length of delivery delay. To examine performance of average inventory level under different system conditions, factors such as ratio of average length of delivery delay to production lead-time, length of production cycle, and coefficient of variation of demand are chosen to define different system conditions. As the performance measure, the ratio of push type inventory to pull type inventory is used in the comparison of the push and pull types, and a quantitative criterion of production model selection is presented. [ABSTRACT FROM AUTHOR]

To deal with the problem that many products in discrete manufacturing factories had flexible process design which had increased the complexity of scheduling, this paper proposed the FPJSP as an extension of JSP, and built the mathematical model for the new scheduling problem to shorten the production cycle. It proposed the hybrid genetic algorithm (HGA) to solve the model. According to the characteristics of FPJSP, it designed the genetic algorithm (GA) with improved chromosome encoding schema and genetic operators. In combination with the variable neighborhood search algorithm (VNS), it designed four neighborhood structures to enable dynamic neighborhood search and improve the neighborhood search performance of GA. The experiment shows the effectiveness and optimize performance of the proposed HGA for solving FPJSP. [ABSTRACT FROM AUTHOR]