4 présentations

• 
  14h45 - 15h07

  An Intelligent Decision Support System for Last-Mile Delivery Optimization under Uncertainty

  • Mahsa Mohammadi, prés., The University of British Columbia
  • Babak Mohamadpour Tosarkani, The University of British Columbia

  This study introduces a multi-stage stochastic dynamic risk-averse model to address the last-mile delivery challenges for online and offline orders. The proposed model incorporates three innovative strategies, including mobile depots, crowd-shippers, and hyper-local drivers. It optimizes mobile depot locations in advance, despite incomplete information about offline orders and driver availability. A deep learning model predicts crowd-shipper availability, utilizing features derived from social exchange theory, while an ensemble model identifies the most critical features. To manage dynamic online orders, a Markov decision process and an online reinforcement-driven adaptive optimization (ORDAO) approach are developed to determine optimal routing policies. The ORDAO method is validated through an interactive multi-agent simulation, which compares average delivery times and the efficiency of delivery tours. The results highlight ORDAO's effectiveness in improving delivery performance and flexibility in dynamic environments.

• 
  15h07 - 15h29

  Cooperation and the Design of Public Goods

  • Juan Carlos Martinez Mori, prés., Georgia Institute of Technology
  • Alejandro Toriello, Georgia Institute of Technology

  We consider the cooperative elements that arise in the design of public goods, such as transportation policies and infrastructure. These design processes involve a wide variety of stakeholders: governments, businesses, advocates, and at the heart of it all transportation users. As such, their eventual deployment is critically dependent on the decision-maker's ability to garner sufficient support from each of these groups. We formalize these strategic requirements from the perspective of cooperative game theory. Specifically, we study non-transferable utility (NTU) linear production (LP) games, which combine the game-theoretic tensions inherent in public decision-making with the modeling flexibility of mathematical programming. We derive conditions under which cooperation is possible, results on the complexity of associated computational tasks, and computational solution methods. To illustrate the practical merits of this theory, we perform a data-driven implementation to study the interplay between various design goals and public reception in the context of public transit networks.

• 
  15h29 - 15h51

  Automated Speed Enforcement Program Interactive Tool for Safety Assessment

  • Mandana Hedayati, prés.,
  • Mehdi Nourinejad, Assistant Professor
  • Ghazaleh Mohseni, PHD Candidate
  • Peter Park, Associate Dean Research Innovation

  Since 2020, numerous municipalities in Ontario have successfully implemented Automated Speed Enforcement (ASE) programs, prioritizing camera installations in areas with severe crash histories. To address the challenge of strategically allocating limited ASE resources, this study presents a GIS-based tool that integrates municipal road networks, demographic data, and historical collision records to identify optimal camera placements and scheduling. Through dynamic Community Safety Zone (CSZ) customization and adjustable safety variable focus ranges, the platform effectively allocates cameras across enforcement cycles, thereby reducing speeding violations. Additionally, a predictive modeling component estimates deterrent effects and citation issuance, which are displayed in real-time charts for policy evaluation. A key feature of this tool is its interactive user interface, enabling planners to relocate camera sites and observe immediate updates in predicted speeding and citation trends. This approach offers a practical, adaptable solution for Ontario municipalities seeking to enhance traffic safety and significantly reduce collisions across jurisdictions.

• 
  15h51 - 16h13

  Integrated Supplier and Third-Party Reverse Logistics Provider Selection in the Laptop Industry

  • Naghmeh Rabiei, prés., Toronto Metropolitan University
  • Saman Hassanzadeh Amin, Toronto Metropolitan University
  • Saeed Zolfaghari, Toronto Metropolitan University

  With growing environmental concerns and stricter regulations, manufacturers must collaborate with suppliers and Third-Party Reverse Logistics Providers (3PRLPs) to enhance waste management and promote the reuse of returned items. 3PRLPs specialize in Reverse Logistics (RL) activities such as recycling, refurbishing, and disassembling, which enable the recovery and reintegration of reusable components into manufacturing. However, selecting suitable suppliers and 3PRLPs is challenging due to varying service quality and performance criteria. This talk addresses the simultaneous selection of suppliers and 3PRLPs in the laptop industry in Canada. Suppliers provide new components, while 3PRLPs process used laptops, recovering and transporting reusable parts back to the manufacturer. Evaluation criteria are categorized as beneficial, target values, and cost. To rank the suppliers and 3PRLPs, two Multi-Criteria Decision-Making (MCDM) methods, Target-based Weighted Aggregated Sum Product Assessment (T-WASPAS) and Target-based MULTIMOORA (T-MULTIMOORA), are employed. Then, the evaluation results for decision-making in sustainable laptop supply chain are discussed.