Abstract:
The construction materials retailing is an interesting and rapid-growth business. Key products that retailers usually focus are ceramic titles and sanitary wares. These products are also bulky fragile and heavy resulted to their high transportation costs. As a part of the logistics planning, the truck allocation planning is focused in this research. A traditional model for truck allocation planning is a deterministic model assuming realization of demands in advance and then computing the optimal solution. The main drawback of this approach is that the model fails to response of any disruption occurs in a distribution network as proven in 2011 Thailand Great Flood, in which the demands were highly fluctuated and trucks required constantly re-allocated responding the situations. To alleviate this drawback, we formulated a truck allocation tactical planning problem in a ceramic distribution network as a two-stage robust optimization model. In the first stage, the model decides numbers and types of trucks in a retailer fleet, whereas the truck allocations and the short term contracts are determined in the second stage. The analysis showed that the model provides more responsive to any disruption as the additional transportation costs resulted of the worst case scenario increases 9.86% compare to that of the average scenario. The analysis also reveals the L-Shaped structure of model; therefore, a heuristic based on the Bender's Decompositions is used to reduce the computation time. The experiment of the heuristic shows the optimal gap is 3.46% and the computation benefit of the heuristic diminishes after 36 hours.