Solving a stochastic inland waterway port management problem using a parallelized hybrid decomposition algorithm

Abstract

This study proposes to develop a mathematical model that captures and appropriately optimizes a number of realistic features (e.g., barge/towboat assignments, maintenance, and availability decisions) for the design and management of an inland waterway transportation network under stochastic commodity supply and water level fluctuations scenarios. To efficiently solve this challenging NP-hard problem, we propose to develop a highly customized parallelized hybrid decomposition algorithm that combines Sample Average Approximation with an enhanced Progressive Hedging and Nested Decomposition algorithm. Computational results indicate that the proposed algorithm is capable of producing high quality solutions consistently within a reasonable amount of time. Finally, a real-life case study is constructed by utilizing the inland waterway transportation network along the Mississippi River. Through multiple experimentations, a number of managerial insights are drawn that magnifies the impact of different key input parameters on the overall inland waterway port operations.