The University of Edinburgh -
Division of Informatics
Forrest Hill & 80 South Bridge

MSc Thesis #9566

Title:Applying Genetic Algorithms to Oil Tank and Pipeline Maintenance
Date: 1995
Abstract:This project applies genetic algorithms to the problem of scheduling activities for the inspection and repair of oil installations such as: tanks and pipelines. A fitness function which evaluates the main aspects of scheduling activities of oil installations maintenance was devised. The constraints considered are: level of production, condition and location of the installations, type of products, human resources and dates of inspection and repairs. Some ideas from similar work of [Langdon 95] and [Clarkson 95] were taken and the implementation was built on Peter Ross' PGA testbed. The data and part of the structures of tanks and pipelines was taken from the Expert System for the Inspection of Tanks and Pipelines SITA and SIGO. This application will be integrated to these two expert systems in order to help inspectors and owners of the installations to give preventive maintenance to oil installations. The implementation was a successful prototype, suggesting that a useful real system can be built in future along these lines.

[Search These Pages] [DAI Home Page] [Comment]