Hybrid dynamic scheduling model for flexible manufacturing system with machine availability and new job arrivals

Scheduling problem in flexible manufacturing system (FMS) is considered dynamic since new orders arrival and machine breakdowns may inevitably render the current schedule inapplicable. This makes rescheduling necessary to instantly handle machine breakdown and accommodate the new arriving orders into the existing schedule while maintaining the efficiency and stability of the current schedule. However, frequent rescheduling may lead to instability and lack of continuity in the existing shop floor schedules. Therefore, this research aims to propose an effective and practical scheduling/rescheduling approach that takes into account the real FMS environment and the desired objectives of manufacturing systems. The proposed approach provides high quality solution with respect to efficiency as well as shop floor stability. The idea of hybridizing the newly developed biogeography based optimization algorithm (BBO) with variable neighborhood structure (VNS) is proposed in order to produce a high performance initial schedule in terms of minimum completion time, tardiness and flow time within reasonable amount of time. Furthermore, due to the limitation of single rescheduling strategy to handle various disruptions, an approach that combines multiple rescheduling strategy is used to maintain efficiency and stability. The hybrid rescheduling strategy takes into account the affected operation rescheduling (AOR) strategy and BBO-VNS match-up approaches to handle machine breakdown and accommodates new arrived order without changing the sequence of operations on machines. The BBO-VNS match-up algorithm manipulates the idle times on machines within the time horizon for assigning the affected operations by breakdown and/or newly arrived orders. Subsequently, a novel approach that combines the hybrid rescheduling strategy with an initial robust schedule which is generated using random fuzzy variables is presented. The aim is to associate an effective hybrid dynamic scheduling model that is able to facilitate the control and accommodation of future disruptions. The performance of the schedules as produced by the scheduling/rescheduling algorithms were investigated and compared. The proposed approaches have been successfully tested on the benchmark test problems and verified in the real FMS scheduling environment based on tardiness and flow time and stability. The statistical analyses demonstrate the efficiency and effectiveness of the proposed hybrid BBO-VNS algorithm over GA, BBO and PBSA to find optimum/near optimum solutions within reasonable amount of time. In addition, the experimental results illustrate the effectiveness of hybrid rescheduling strategy for handling complex disruptions, in which schedules with high quality stable are produced. On average, the hybrid rescheduling approach improves the performance with respect to both the average efficiency measure (AEM) and average stability measure (ASM) obtained under total rescheduling (TR) by 17.07%, AOR by 5.58%,route change rescheduling (RCR) by 9.75%, right-shift rescheduling (RSR) by 25.50% and BBO-VNS match-up by 4.01%. Furthermore, the results of combined hybrid rescheduling strategy with initial robust schedule in presence of multiple disruptions confirmed that their combination is effective in which even more reliable high quality stable schedules are delivered in which the combined approach achieves significant improvement over hybrid rescheduling strategy based on the initial predictive schedule strategy is 5.13%, over modified AOR is 16.19% and over TR is 19.43% with respect to the mean of efficiency and stability. Therefore, the experimental results have favorably shown that the proposed improved hybrid dynamic scheduling is effective and practical in providing reliable solutions in a real world dynamic and uncertain FMS with respect to desired objectives of manufacturing system.

Preview PDF FK 2015 100RR.pdf Download (1MB) | Preview

Actions (login required)

View Item