A Tabu Search Approach for Minimizing Fuel Consumption on Cyclic Natural Gas Pipeline Systems

Abstract: In this paper we propose a tabu search (TS) heuristic for the problem of minimizing fuel usage on cyclic natural gas transmission networks in steady-state. The problem is modeled as a nonconvex nonlinear program. Although effective approaches, mainly based on dynamic programming, exist for handling non-cyclic topologies, the cyclic case has not been addressed satisfactorily. In this work, we present a successful short-term memory strategy that overcomes local optimality, and provide empirical evidence, based on a number of instances using real-world data, of the superiority of the proposed TS over existing approaches.