A novel matheuristic based on bi-level optimization for the multi-objective design of hydrogen supply chains

This work introduces an efficient tool for the design of sustainable hydrogen supply chains (HSCs), considering both economic and environmental concerns, through an appropriate multi-objective strategy. The original problem, being formulated as a bi-objective mixed-integer linear programming (MILP) problem, takes into consideration the availability of different energy sources, the installation and operation of hydrogen facilities of different sizes and technologies, and the transportation of hydrogen from production units to storage facilities. The area of study is divided into grids which have a specific hydrogen demand that evolves over time, thus a multi-period model of the HSC is considered. In order to overcome the computational burden associated to the solution of large size instances of the resulting problem, we proposed a solution strategy consisting of a hybrid algorithm. The original problem is reformulated into a bi-level optimization problem: the upper level (discrete problem) consists of finding the optimal location for production plants and storage facilities, whereas the lower level (continuous problem) minimizes their corresponding costs associated to transportation and facility operation. A multi-objective evolutionary algorithm is employed for the solution of the bi-objective upper level, whereas the bi-objective lower level is decomposed using a scalarizing function, which is then solved using a linear programming solver. The proposed methodology is validated through the comparison of the true Pareto fronts given by CPLEX with -constraint method, for six increasing size instances. Numerical results prove that the proposed hybrid approach produces an accurate approximation of the Pareto-optimal fronts, more efficiently than the exact solution approach.Peer ReviewedPostprint (author's final draft