This work presents a novel hybrid algorithm called GA-RRHC based on genetic algorithms (GAs) and a random-restart hill-climbing (RRHC) algorithm for the optimization of the flexible job shop scheduling problem (FJSSP) with high flexibility (where every operation can be completed by a high number of machines). In particular, different GA crossover and simple mutation operators are used with a cellular automata (CA)-inspired neighborhood to perform global search. This method is refined with a local search based on RRHC, making computational implementation easy. The novel point is obtained by applying the CA-type neighborhood and hybridizing the aforementioned two techniques in the GA-RRHC, which is simple to understand and implement. The GA-RRHC is tested by taking four banks of experiments widely used in the literature and comparing their results with six recent algorithms using relative percentage deviation (RPD) and Friedman tests. The experiments demonstrate that the GA-RRHC is a competitive method compared with other recent algorithms for instances of the FJSSP with high flexibility. The GA-RRHC was implemented in Matlab and is available on Github.
The Flexible Job Shop Scheduling Problem (FJSSP) continues to be studied extensively to test new metaheuristics and because of its closeness to current production systems. A variant of the FJSSP uses fuzzy processing times instead of fixed times. This paper proposes a new algorithm for FJSSP with fuzzy processing times called the global neighborhood with hill-climbing algorithm (GN-HC). This algorithm performs solution exploration using simple operators concurrently for global search neighborhood handling. For local search, random restart hill-climbing is applied at each solution to find the best machine for each operation. For the selection of operations in hill climbing, a record of the operations defining the fuzzy makespan is employed to use them as a critical path. Finally, an estimation of the crisp makespan with the longest processing times in hill climbing is made to improve the speed of the GN-HC. The GN-HC is compared with other recently proposed methods recognized for their excellent performance, using 6 FJSSP instances with fuzzy times. The obtained results show satisfactory competitiveness for GN-HC compared to state-of-the-art algorithms. The GN-HC implementation was performed in Matlab and can be found on GitHub (check Data Availability Statement at the end of the paper).
Este artículo aborda la programación de tareas en el Flexible Job Shop Scheduling Problem (FJSSP). En este sistema de manufactura es necesario incrementar el número de trabajos a procesar debido a las condiciones actuales del sector industrial en donde existe un aumento en la demanda de productos, lo que conlleva a incrementar la producción. Para encontrar una programación de tareas cercana al óptimo. Se propone un método de optimización híbrida utilizando una búsqueda global basada en algoritmos genéticos (AG) que tienen buena diversificación y para la búsqueda local se aplica una escalada de colinas simple con reinicio (ECR) para mejorar cada solución. La combinación de estas metaheurísticas obtiene el equilibrio necesario para encontrar la mejor programación de tareas con el fin de minimizar el makespan como función costo. Se implementó el algoritmo propuesto en Matlab, para comprobar su eficiencia se compararon los resultados con investigaciones recientemente publicadas.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.