Bi-objective optimization of forest-based biomass supply chains for minimization of costs and deviations from safety stock

Abstract

• Developed a bi-objective optimization model for forest-based biomass supply chain. • Incorporated practical considerations related to biomass safety stock in the optimization model. • Minimized the total cost and the deviations from the desired safety stock and derived their tradeoffs. • Applied the bi-objective model to a real case of producing syngas at a pulp mill. • Cost savings could be achieved by allowing deviations from the desired safety stock. Bioenergy from forest-based biomass could reduce fossil fuel dependency, GHG emissions, and generated wastes. However, the inherent complexity, uncertainty, and high cost of forest-based biomass supply chains could hinder the economic viability of bioenergy projects. The supply chain activities including transportation, storage, handling, and preprocessing of biomass are interdependent and require proper planning. Therefore, in previous studies, mathematical programming models were developed to plan and optimize the biomass supply chain activities. In this paper, a bi-objective optimization model is developed for tactical planning of the forest-based biomass supply chains in order to determine the trade-offs between the total costs and the possible deviations from the safety stock. The first objective is to minimize the upstream supply chain costs, and the second objective is to minimize the negative deviations of monthly inventory from the safety stock. The decision variables include the optimal monthly biomass flows, preprocessing, and inventory levels. The model is applied to the case of a biomass gasification at a Kraft pulp mill in British Columbia, Canada. The output of the model is a set of Pareto optimal solutions, which shows the trade-offs between the objectives of cost and safety stock deviation. This approach has the potential to assist the decision makers by providing a set of solutions to opt from based on their preferences and their attitude towards supply disruption risk. The results indicate a maximum of 18% cost savings is possible if the inventory level deviates from the safety stock. A sensitivity analysis is also performed to assess the impact of variations in the biomass availability and cost, and feedstock demand on the Pareto optimal solutions. According to the sensitivity results, feedstock demand is the most sensitive parameter.