Controllable processing times with limited resources and energy consumption in flow shops: an assessment by simulation

In production systems, the total energy available is a constraint due to energy costs or power consumption. The control of the processing time can affect the total energy consumed for the task performed. This paper proposes two control policies to change the processing time of stations that composes a flow line under resource consumption constraint in a production period. The first reduces the processing time of one station (centralized), while the second reduces the processing time of more stations (distributed). The introduction of an accumulator of the resources (such as a battery) can use any residual energy of a period in the following period. A multi-agent architecture has been developed to support the proposed models. This architecture decomposes the main problem into several sub-tasks, reducing the computational complexity and enabling the solution of larger and industrial problems. The simulation results highlight how the centralized approach works better when fewer resources are available, and the production line is unbalanced. On the other hand, the distributed model is more suitable for a balanced flow line. The introduction of an accumulator improves the performance of the centralized model compared to the distributed model.