Conference Agenda

Disruptions, such as unexpected incidents, pose irreparable costs to the supply chain (SC) network. In this paper, to protect against these disruptions, a robust three-objective optimization model is developed for a resilient and responsive SC under uncertainty. Objectives are costs, quality of the raw and final products, and response time in designing a four-level resilient and responsive SC, including suppliers, manufacturers, distributors, and customers. Resilient strategies are employed to manage the possible disruption for suppliers and manufacturers, including repair and improvement, strengthening, and backup sourcing. In the first place, if the disruption occurs to the suppliers, backup suppliers should be replaced. However, this replacement requires extra costs, and the backup suppliers may not have the essential quality. Therefore, in the developed model, the best supplier is chosen based on its raw material quality, resilience ability, and cost. Secondly, if the disruption happens to the manufacturers, in addition to the backup manufacturer option, another resilience strategy can be employed, named repair and improvement. In this strategy, a penalty cost should be paid proportionate to the time spent repairing, and the current manufacturer should continue producing the ordered product. Additionally, in this study, the operational uncertainty of some parameters is dealt with by utilizing robust optimization. Finally, the Benders decomposition algorithm is proposed to solve the NP-hard model in large sizes.

In recent years, customers have been rediscovering local markets while firms seek ways to shorten supply chains and strengthen local industries. With increasing interest in local and regional produce, mobile stores, known as pedlars, are gaining attention as a channel for supporting local entrepreneurs. This study addresses the challenge of selecting optimal locations for mobile stores when customer demand, availability, and spending at locations are uncertain. We evaluate various location selection policies using agent-based simulation and compare them to a reinforcement learning approach. To this end, we model and develop policies based on practitioner interviews. Using agent-based simulation allows us to model and simulate seemingly random differences between locations as described by practitioners, thus providing decision support.

In this study, we have designed a multiobjective decision support tool prototype for robust forest harvest scheduling in multiple periods. Indeed, we formulated and solved a novel multi-scenario multiobjective mixed-integer optimization problem, with 108 objective functions, providing support for forest planners to study the existing trade-offs between demand satisfactions for multiple assortments in different planning periods. Besides, an interactive robust analysis to study the outcomes variations caused by uncertainty and find a robust schedule in tactical forest planning problems has been proposed. We validated the useability of the proposed decision-support prototype with a Swedish case study with a harvest planning horizon of twelve months.