Dynamic AGV-Container Job Deployment Strategy

Automated Guided Vehicles (AGVs) are now becoming popular in container-handling applications at seaport. Efficacy of the dispatching strategy adopted to deploy AGVs is a prime factor affecting the performance of the entire system. The objective of this project is thus to develop an efficient dispatching strategy to deploy AGVs in a container terminal. The scenario considered was a container terminal where containers are uploaded to and discharged from ships. Discharged containers are stored at specific storage locations in the terminal yard. Containers are moved between dock and yard by a dedicated fleet of AGVs. At any point of time, each AGV carries at most two containers. This two-container load may comprise of any plausible permutation of containers for discharge or upload. To reduce congestion and increase utility level, an efficient dispatching strategy for AGVs is paramount. At present, a variety of heuristic methods for dispatching AGVs are available, but these methods were primarily developed to work in a manufacturing context where the network structure is uncomplicated and only a small number of AGVs are required. The situation under consideration entails greater network complexity and also a large fleet of close to 80 AGVs. In this study, the problem was modeled via network flows with constraints, which describe the disparate instances when the AGV carries one container and when it carries two. Heuristic algorithms based on this model are proposed and their performance investigated.Singapore-MIT Alliance (SMA