Global Search Metaheuristics for planning transportation of multiple petroleum products in a multi-pipeline system
Abstract
The objective of this work is to develop several metaheuristic algorithms to improve the efficiency of the MILP algorithm used for planning transportation of multiple petroleum products in a multi-pipeline system. The problem involves planning the optimal sequence of products assigned to each new package pumped through each polyduct of the network in order to meet product demands at each destination node before the end of the planning horizon. All the proposed metaheuristics are combinations of improvement methods applied to solutions resulting from different construction heuristics. These improvements are performed by searching the neighborhoods generated around the current solution by different Global Search Metaheuristics: Multi-Start Search, Variable Neighborhood Search, Taboo Search and Simulated Annealing. Numerical examples are solved in order to show the performance of these metaheuristics against a standard commercial solver using MILP. Results demonstrate how these metaheuristics are able to reach better solutions in much lower computational time.
