Combining matheuristics and MILP to solve the accessibility windows assembly line balancing problem level 2 (AWALBP-L2)

An outer approximation method for an integration of supply chain network designing and assembly line balancing under uncertainty

“…Avikal, Jain, Mishra, and Yadav (2013) addressed the U shaped assembly line balancing problem with consideration of labor productivity. Calleja, Corominas, García-Villoria, and Pastor (2014) addressed the socalled Accessibility Windows Assembly Line Balancing Problem (AWALBP) where the length of the work-piece is larger than the width of workstations. Ullah, Guan, Liu, Zhang, and Wang (2014) studied the assembly line balancing problem with uncertain task times.…”

Hybrid metaheuristics for the Accessibility Windows Assembly Line Balancing Problem Level 2 (AWALBP-L2)

“…and instances up to a certain size were solved to optimality. On the other hand, a different approach to the problem considers hybridizing heuristics and mathematical programming to solve the instances that are out of reach of the former models, which is presented in Calleja et al (2014). In that work, a heuristic is used to compute good initial solutions and to derive bounds.…”

“…In a second step, the obtained bounds are incorporated onto a MILP model, with the aim to reduce the model dimension and the solution space. Contrarily to the collaborative combination between a heuristic and MILP of Calleja et al (2014), in this paper we propose integrative combinations between metaheuristics and MILP, where the MILP model is a subordinated element embedded inside the metaheuristic scheme.…”

“…What emerges from the computational results on AWALBP-L2 (Calleja et al, 2013(Calleja et al, , 2014 is that computing an optimal solution of the problem might become intractable for large size instances. This may be expected since even the simplest version of the problem, AWALBP-L1, is already NP-hard, as proven in Calleja et al (2014). For this reason, metaheuristic or hybrid solution methods could be envisioned to solve this problem.…”

“…The two first hybrid metaheuristics proposed in this work, SA and TS, rely on a mathematical programming model: the so-called Task model of Calleja et al (2014). The Task model computes, for a given movement scheme, the optimal assignment of each task to one stationary stage of the cycle.…”