Abstract
This paper proposes a Lagrangian relaxation approach toward the solution of the multistage lot-sizing problem with multiple constrained resources in general product structures. Dualizing the resource constraints this problem results in an uncapacitated problem that can be solved efficiently by reformulating it as a simple facility location representation. The Lagrangian dual costs are updated by subgradient optimization. A Lagrangian based heuristic procedure to generate good solutions for the problem is also proposed. Numerical results are presented for some test problem instances of differing sizes.
(Lot-sizing; Capacity constraints; Echelon stocks; Lagrangian relaxation)
