Joint optimization of demand-side operational utility and manufacture-side energy consumption in a distributed parallel machine environment

Previous production scheduling models often set optimization objectives from the perspective of manufacturers, such as makespan, tardiness and energy consumption. However, none of the objectives can reflect the extent to which the scheduling plan affects the demand side. In fact, the delivery time of orders will directly affect the equipment utilization or project schedule on the demand side. In this paper, we focus on a new objective named total operational utility of all distributed equipment from the demand side, and integrate it into an energy-efficient production scheduling model based on the distributed parallel machine environment, in which the total energy consumption of manufacture side including processing energy consumption and transportation energy consumption is another objective. The orders are the spare parts used to replace the deteriorated components of distributed equipment based on forecasting information. Based on the scheduled delivery time fed back from the scheduling plan, the relationship among operating speed, deterioration rate and operating efficiency is used, and an optimal speed adjustment strategy is developed for each equipment to improve the operational utility. A memetic algorithm (NMA) based on the structure of NSGA-II is presented for the model. A list scheduling heuristic and a problem-dependent heuristic are designed to generate initial population. Two problem-dependent local search operators are developed to enhance the searching ability. By performing extensive experiments and comparing NMA with some well-known algorithms, the effectiveness and superiority of NMA are demonstrated.

Automated summary

This paper proposes a production scheduling model that takes into account the operational utility from the demand side and integrates energy efficiency goals. By optimizing the operating speed of equipment based on forecasting information and delivery time, the operational utility is improved. Experimental results demonstrate the superiority of the model.