ABSTRACT
This paper aims to quantify the effects of production disruptions (PDs) and physical distancing constraints due to the pandemic in a parallel-machine production environment. The machines are non-identical and are utilized for producing a finite set of jobs (parts) in a plastic injection moulding production. The production process is subjected to random production downtime disruptions. A mixed-integer linear programming (MILP) model is developed for optimizing the joint production plan and schedule, which maximizes the production's total benefit. The model is utilized to plan and schedule a set of 17 machines in a Canadian manufacturing company. To explore the effects of physical distancing and PDs on the production's total net profit, four different scenarios for normal operation and production during the pandemic, with and without production downtimes, are considered. A genetic algorithm is utilized to solve the model. The results show that considering machines' random breakdowns and physical distancing individually reduces the total profit of the production by 71.58 and 57.98%, respectively;while their joint effect results in a 88.54% reduction in the annual net profit.
ABSTRACT
This paper, for the first time, presents a production scheduling model for a production line considering physical distancing between the machines' workforces. The production environment is an unrelated parallel-machine, in which for producing each part, different machines with different production rates and the required number of workers are available. We propose a three-objective mixed-integer linear programming mathematical model that aims to maximize the manufacturer's total benefit, parts' safety stock (SS) index, and the workforce's physical distance over a finite horizon (one year) by determining the optimal scheduling of the parts on the machines. Since a large production scheduling problem belongs to the Np-Hard category of problems, a non-dominated sorting genetic algorithm, and a non-dominated ranked GA algorithm are developed to solve the presented model in two stages using the empirical data from a Canadian plastic injection mold company. In the first stage, the LP-metrics approach is utilized for validating the meta-heuristics on a reduced-size problem. In the second stage, the validated meta-heuristics are utilized to optimize the company's yearly production schedule. The results indicate both metaheuristics are performing well in determining the optimal solution. Moreover, implementing physical distancing in the company reduces the company's monthly net benefit by around 9.56% compared to the normal operational conditions (without considering physical distancing).