We present an application of a multi-agent cooperative search approach to the problem of optimizing gas pipeline operations, i.e.\ finding control parameters for a gas transmission network that result in a low usage of energy to make the required gas deliveries. Our cooperative search approach improves on the pure competition of search agents by having them exchange good solutions from time to time that both are integrated into the search state of the agents and used to improve the search control of the agents. Our experimental evaluation with real problem instances from TransCanada show that our system meets TransCanada's time requirements and reliably outperforms the interactive method that is the current state-of-the-art by creating solutions that require more than 10 percent less energy.
Generated: 25/05/2008