Search

This paper uses a Genetic Algorithm (GA) to reduce total tardiness in an identical parallel machine scheduling problem. The proposed GA is a crossover-free (vegetative reproduction) GA but used for four types of mutations (Two Genes Exchange mutation, Number of Jobs mutation, Flip Ends mutation, and Flip Middle mutation) to make the required balance between the exploration and exploitation functions of the crossover and mutation operators. The results showed that use of these strategies positively affects the accuracy and robustness of the proposed GA in minimizing the total tardiness. The results of the proposed GA are compared to the mathematical model in terms of the time required to tackle the proposed problem. The findings illustrate the ability of the propounded GA to acquire the results in a short time compared to the mathematical model. On the other hand, increasing the number of machines degraded the performance of the proposed GA. [ABSTRACT FROM AUTHOR]

Purpose: This paper studies a 3-dimensional pallet loading problem considering interlock stacking pattern, box dimensions, humidity, and storage time, where overlapping and overhanging are not allowed. Despite the importance of this problem in the literature, our work provides the first method that considers the environmental conditions such as 1) storage time and 2) humidity, and their tremendous impacts on the strength of the boxes, as has been observed widely in the DHL supply chain. Design/methodology/approach: This paper proposes a two-phase heuristic algorithm to solve a 3-dimensional pallet loading problem under real conditions (relative humidity, and storage time) considering interlock stacking patterns, where overlapping and overhanging are not allowed. In phase 1, the horizontal layer configuration is determined by block techniques. Three types of horizontal layers are created based on box dimensions perpendicular to the base. In phase 2, a novel mathematical model is propounded to improve the pallet volume utilization, and stability considering the pallet's maximum allowable height and weight, and the dynamic compression strength of boxes. The dynamic compression strength of boxes is calculated by the modified McKee formula. Two performance measures, pallet volume utilization, and stability (load height) are utilized to evaluate the performance of the proposed heuristic algorithm in real-world instances (DHL Supply Chain). Findings: The results illustrated that the dynamic compression strength of boxes decreases as the relative humidity and storage time increase. The load height changes dynamically along with box dimensions, box alignment, direction, relative humidity, and storage time. Increasing relative humidity and storage time and applying an interlock stacking pattern reduce the pallet utilization, however, enhance the pallet stability. Finally, the proposed heuristic algorithm's efficacy increases as the identical box dimensions' heterogeneity increases. Originality/value: It is believed in the supply chain where these characteristics are observed, the implementation of the heuristic algorithm will help them improve the pallet volume utilization and stability. [ABSTRACT FROM AUTHOR]

This paper aims to study tasks' allocation problem in a single-station collaborative robot assembly system 4.0. Two products with identical demands are run in a system simultaneously in mixed mode to reduce the idle time and improve the system utilization. The idle time occurs due to balancing difficulties and particularly in this case restrictions in the task vs. robot-worker feasibility matrix. It is not allowed for the resources to assemble a same product simultaneously to avoid any direct contact between resources. Thus, the products are swapped between resources as needed to complete the assembly tasks. Two approaches, a novel mathematical model and the shortest processing time (SPT) rule, are propounded in this study to reduce the cycle and idle times and enhance the production rate of an assembly system. The two proposed methods rely on dividing the timeline of a station into unknown stages with unknown lengths. The performance of the proposed mathematical model and SPT rule are evaluated by two performance measures: the total number of stations and the % average station utilization. The obtained results are compared to the results of the modified COMSOAL heuristic. The findings revealed that the novel mathematical model dominated results. It guaranteed minimizing the cycle time and maximizing the production rate compared to the SPT rule and modified COMSOAL heuristic. However, many companies value a 1-s reduction in cycle time as valuable. This directly relates to the reduction of direct resources costs and the total number of stations. [ABSTRACT FROM AUTHOR]