A direct search algorithm for solving the multi-period single-sourcing problem

Abstract

Generally, problems in logistics faced by a supplier would be the production timing, the location of inventories, and the assignment of customers to warehouses. This paper will consider a dynamic multi-period single-sourcing problem (MPSSP) used to support the corresponding decisions. We propose a direct search algorithm for solving the multi-period single-sourcing problems. In partiular, we generalize the strategy of releasing nonbasic variables from their bounds, combined with the active constraint method that was developed for the Generalized Assignment Problem (GAP) to a class of convex assignment problems. We then identify an important subclass of problems, containing many variants of the multi-period single-sourcing problem (MPSSP), as well as variants of the GAP.