Enhanced memetic search for reducing energy consumption in fuzzy flexible job shops

Abstract

The flexible job shop is a well-known scheduling problem that has historically attracted much research attention both because of its computational complexity and its importance in manufacturing and engineering processes. Here we consider a variant of the problem where uncertainty in operation processing times is modeled using triangular fuzzy numbers. Our objective is to minimize the total energy consumption, which combines the energy required by resources when they are actively processing an operation and the energy consumed by these resources simply for being switched on. To solve this NP-Hard problem, we propose a memetic algorithm, a hybrid metaheuristic method that combines global search with local search. Our focus has been on obtaining an efficient method, capable of obtaining similar solutions quality-wise to the state of the art using a reduced amount of time. To assess the performance of our algorithm, we present an extensive experimental analysis that compares it with previous proposals and evaluates the effect on the search of its different components.

A weakly supervised active learning framework for non-intrusive load monitoring

“…Energy efficiency has gained great traction in recent years [1][2][3], to facilitate the transition to NETzero economy. Energy awareness plays a key role in improving energy efficiency [4][5][6][7][8][9], and active user participation can potentially increase a household's energy flexibility, leading to energy savings of up to 30% [8].…”

“…Moreover, the findings of the study highlight the need to develop strategies and technologies that are more user-centred. Energy awareness can be improved by monitoring energy [3] and particularly via Load Monitoring that provides detailed information about consumption. Specifically, Non-Intrusive Load Monitoring (NILM) is a purely algorithmic approach to estimate individual appliance power consumption that contribute to the measured aggregate signal, via 1 This is a peer-reviewed, accepted author manuscript of the following research article: Tanoni, G, Sobot, T, Principi, E, Stankovic, V, Stankovic, L & Squartini, S 2024, 'A weakly supervised active learning framework for non-intrusive load monitoring', Integrated Computer-Aided Engineering.…”

Minimizing the makespan and carbon emissions in the green flexible job shop scheduling problem with learning effects

“…They proposed an energy-saving strategy to reduce the machine speed for noncritical work without changing the makespan. Later, García Gómez et al 13 focused on minimizing energy consumption for the FJSP and developed a memetic algorithm that combines the global search with local search for model solving. Current studies on job shop scheduling problems assume that the processing time from the first job to the last one is known and fixed.…”