Container-ships are vessels possessing an internal structure that facilitates the handling of containerised cargo. At each port along the journey of a container-ship, containers destined for that port are unloaded, and some containers destined for subsequent ports are loaded. Determining a viable configuration of containers that facilitates this unloading and loading, in a cost-effective way, constitutes the deepsea container-ship stowage problem. The work of determining a stowage configuration for a container-ship, on leaving a port, is performed by human stowage planners. The success of a configuration requires consideration of many factors. These factors include whether the configuration contravenes ship stability, minimises the physical costs of handling the containers, and takes into account expected container loads at subsequent ports. Further complications arise from the existence of hazardous cargo that must be segregated from other cargo and the ship's crew, and from the need to handle containers of non-standard dimensions. Stowage planners must work under strict time constraints, and are limited in the number of stowage configurations that they can consider. This real-world problem seems to be one that would benefit from automation through the application of artificial intelligence. Although many decision support systems exist that automate the time-consuming calculations for ship stability, little work has been published in the area of full automation of stowage planning. Authors proposing full automation have correctly identified the salient features of the problem, but have allowed the array-like nature of spaces within containerised vessels to entirely dictate their approach to addressing the placements of specific containers to specific locations. To enable the implementation of these approaches, excessively large search spaces are pruned through the removal of important features of the problem, rendering the solutions not commercially viable. By concentrating solely on the specific placements of containers, these authors have not recognised how human planners solve the problem. The author of this thesis approaches the container-ship stowage problem from a knowledge engineer's perspective. In the proposed approach, 'intelligence' is provided through the application of the findings of a knowledge elicitation exercise and a systems analysis of the work of human planners. The assumed heuristics inherent in their use of documents are highlighted. This thesis reports on the results of the analysis of the processes employed by a stowage planner. Explanations are provided of how these results allow the problem to be decomposed into subproblems. An implementation of the approach described would determine good, although not necessarily optimum, solutions to the entire problem in a commercially viable duration of time. Further, this approach allows many more stowage configurations to be considered than would be possible for a human planner. The work contained within this thesis demonstrates the feasibility and benefits of such an implementation. The last chapter contains, in addition to a full and detailed list of conclusions made during the research, a summary of some of those areas that still require further work.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:265471 |
| Date | January 1997 |
| Creators | Wilson, Ian David |
| Contributors | Roach, Paul |
| Publisher | University of South Wales |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | https://pure.southwales.ac.uk/en/studentthesis/the-application-of-artificial-intelligence-techniques-to-the-deepsea-containership-cargo-stowage-problem(3b7b5338-5b76-4fe6-8a61-e9704940d23f).html |
Page generated in 0.1431 seconds