3 présentations

• 
  15h30 - 15h55

  A robust optimization approach to emergency response network design for railroad transportation of hazardous materials

  • Manish Verma, prés., DeGroote School of Business, McMaster University
  • Jyotirmoy Dalal, Sheffield University

  For rail-hazmat incident response planning problem, accessing detailed information about location, volume and incident type would be extremely challenging. In addition, dearth of historical data might undermine the long-term viability of strategic decisions about facilities and equipment. We present three robust two-stage stochastic programs to study uncertainty in the objective function and the constraints: first considers uncertainty in the cost of inadequate response; second considers uncertainty in the hazmat volume released; and third considers uncertainty in both inadequate response cost and hazmat volume. The proposed optimization programs are used to study a realistic size network in Canada, and the results are analyzed to generate managerial insights and to inform appropriate public policy decisions.
• 
  15h55 - 16h20

  Optimizing Sustainable Containerized Multimodal Transportation with Resilience and Cost Reduction Strategies

  • Asefeh Hassani Goodarzi, prés., École de technologie supérieure ÉTS
  • Armin Jabbarzadeh, Département de génie des systems, École de technologie supérieure
  • Marc Paquet, École de technologie supérieure

  Containerized multimodal transportation, as a sustainable transport, is an important measure for combatting global warming by combining road transport with more sustainable modes such as railways and waterways. This study evaluates the cost and environmental impact of these modes, in terms of greenhouse gas emissions, as well as externalities such as noise pollution, congestion, and accidents, to ensure social sustainability. To enhance the resilience of the multimodal network against disruptions, a two-stage mixed integer mathematical formulation is presented that integrates consolidation and resilience concepts. Different disruption scenarios are generated based on a case study of the United Kingdom transportation system, and the model is solved using a Lagrangian relaxation method to identify critical and semi-critical routes to be fortified by preparedness strategies, as well as the flow of containers in the system. Numerical results show that cost savings can be achieved by implementing preparedness and recovery actions and consolidation strategies, as follows: By allocating a budget of 0.3% to 0.4% of the total cost towards preparedness and recovery actions, cost reductions of 3% to 4.7% can be achieved. Additionally, implementing consolidation operations can result in savings of approximately 34 times the budget spent.
  Keywords: Containerized multimodal transportation: Sustainable transport: Social sustainability; Resilience; Consolidation.
• 
  16h20 - 16h45

  Managing disruption of intermodal-hazmat shipments through optimization and machine learning

  • Manish Verma, prés., DeGroote School of Business, McMaster University
  • Arash Rad, McMaster University
  • Atiq Siddiqui, Imam Abdulrahim Bin Faisal University

  Rail-truck intermodal networks serve as major freight infrastructure, transporting both regular and hazardous material. Disruptions in its service legs or at intermodal terminals due to accidents, infrastructure failures, etc., cause a significant increase in operational cost and the risk to environment, property and life. These can be mitigated by understanding the criticality of transportation infrastructure components and then developing strategies to offset the adverse impacts. We propose an optimization-machine learning methodology that enables us to categorize the infrastructure based on impact levels, which informs the development of appropriate mitigation strategies. The proposed methodology was applied to a realistic rail-truck intermodal network in the United States and to conclude that: post-disruption consideration should be incorporated in the transportation planning problem; machine learning algorithms can efficiently categorize network elements with a high degree of accuracy; and efficient pro-active post-disruption management can avoid a significant increase in either cost or the associated risks.