Bound Improvement for LNG Inventory Routing

Liquefied Natural Gas (LNG) is steadily becoming a common mode for commercializing natural gas. In this paper, we develop methods for improving both lower and upper bounds for a previously stated form of an LNG inventory routing problem. A Dantzig-Wolfe-based decomposition approach is developed for LNG inventory routing problem (LNG-IRP) attempting to overcome poor lower bounds. However, it fails to find feasible integer solutions that would provide good upper bounds. Advanced construction heuristics based on greedy randomized adaptive search procedure (GRASP) and rolling-time windows and several novel, MIP-based neighborhood search techniques are developed to achieve improved solutions in shorter computational time. The proposed algorithms are evaluated based on a set of realistic test instances that are very large relative to most of the problem instances seen in recent literature. Extensive computational results indicate that the proposed methods find improved lower bounds significantly faster than commercial solvers, and further, optimal or near optimal feasible solutions are achievable substantially faster than commercial optimization software as well as previously proposed heuristic methods.