A new hybrid method for buffer sizing and machine allocation in unreliable production and assembly lines with general distribution time-dependent parameters

Abstract

This paper proposes a multi-objective mathematical formulation and a hybrid approach to solve buffer sizing and machine allocation problems simultaneously in unreliable production and assembly lines. This paper unlike prior researches assumes that time-dependent parameters of production systems are generally distributed (e.g., uniform, normal, gamma, etc.) and not only deterministic or exponential. This paper proposes a multi-objective mixed binary integer nonlinear mathematical model to solve the problem of buffer sizing and machine allocation. The proposed mathematical model is capable of purchasing new machines (candidate) and also selling old machines (current available). In other words, this model compares the candidate machines to current available machines in each station based on different aspects and is capable to replace the current machines with candidate machines or to sell some of the current machines without replacement. To solve the mentioned problem, a new formulation for dealing with multi-objectiveness of the problem is proposed. This formulation generates a series of nondominated solutions, and also, it is capable of generating a non-dominated solution between two adjacent non-dominated solutions determined by decision maker. A hybrid genetic algorithm (HGA) with a new dynamic mutation probability is proposed to solve the model. Since the proposed mathematical model and the proposed solution method are novel, the proposed HGA is compared to simple genetic algorithm and non-dominated sorting genetic algorithm (NSGA-II). The computational results indicate the effectiveness of the proposed HGA.

Scheduling trucks in cross docking systems with temporary storage and repetitive pattern for shipping trucks

“…The quantity of products in inbound trucks is as follows according to the format inbound truck: [(product type, quantity). 1: [(1, 53), (2,41), (3,24), (4,31), (5,35)]; 2: [(1, 54), (2,39), (3,25), (4,34), (5,37)]; 3: [(1, 53), (2,39), (3,25), (4,33), (5,34)]. …”

“…Therefore, the mutation operator increases the possibility of obtaining the global solution [25]. The Mutation rate (P m ) indicates the probability that a chromosome will undergo mutation.…”

“…The quantity of products in inbound trucks is as follows according to the format inbound truck: [(product type, quantity). 1: [ (1,11), (2,12), (3,12), (5,10), (6,13)]; 2: [ (1,12), (2,14), (3,16), (4,15), (5,10), (6,12)]; 3: [(1, 10), (2,11), (3,15), (5,12), (6,18)]; 4: [ (1,13), (2,15), (3,17), (4,25), (5,10), (6,11)]; 5: [ (1,11), (2,14), (3,18), (4,11), …”

“…Genetic Algorithms have proven to be very adaptable to a great variety of different complex optimization tasks, and many researches apply this algorithm to solve the different types of problems [23][24][25][26][27][28][29][30][31]. The main role of genetic algorithm (GA) is to perform global search within all feasible search spaces.…”

Enfoque híbrido metaheurístico AG-RS para el problema de asignación del buffer que minimiza el inventario en proceso en líneas de producción abiertas en serie

“…Para la mejor exploración del espacio de soluciones del BAP, una propuesta reciente es la hibridación de técnicas metaheurísticas con otros métodos tales como: Particiones Anidadas (Shi and Men, 2003), métodos de Ramificación y Acotamiento (Dolgui et al, 2007), Método del Subgradiente (Demir and Tunali, 2008), y Búsqueda Local (Amiri and Mohtashami, 2012), (Mohtashami, 2014). Existen también estudios que reportan la hibridación entre técnicas metaheurísticas, (Kose and Kilincci, 2015) hibridan Algoritmos Genéticos con Recocido Simulado con el fin de maximizar la tasa promedio de producción; por su parte (Su et al, 2017) resuelven el BAP bajo un criterio multiobjetivo en donde se busca maximizar la tasa promedio de producción a la vez que se minimiza el total de buffers asignados, se aplica la hibridación de Búsqueda Tabú y NSGA-II; finalmente (Kose and Kilincci, 2018) resuelven el problema bajo un criterio multiobjetivo similar al empleado en el estudio anterior, a través de hibridar NSGA-II y Recocido Simulado.…”