An inventory-routing problem is one of the combinatorial optimization problems that are very
difficult to solve optimally. The inventory-routing system considered is composed of a central
warehouse and a set of suppliers who produce non-identical items. The items are kept at the central
warehouse and face constant and deterministic demands. An inventory policy with fixed order quantity
is adopted. When a group of items need to be replenished, a capacitated vehicle is dispatched from the
central warehouse to collect them at suppliers' locations. In this paper, constructive heuristics and an
improvement algorithm called a very large scale neighborhood (VLSN) search algorithm are
presented. The constructive heuristics sequentially generate an initial feasible solution. Then, a VLSN
search algorithm is applied to improve the solution by searching very large neighborhoods.
Computational results indicate that the VLSN search algorithm can provide near-optimal solutions
very efficiently.