4 Presentations

• 
  01:00 PM - 01:22 PM

  Subscription and Per-Order Pricing Programs in Temporally-Consolidated Last-Mile Delivery

  • Behnaz Naeimian, presenter, York University
  • Opher Baron, Rotman School of Management
  • Mehdi Nourinejad, York University
  • Peter Park, York University

  This study introduces a Temporally-Consolidated (TC) delivery service aimed at improving the viability of e-commerce in low-density areas by lowering costs and boosting profitability. Through an analysis of subscription and per-order pricing, customer utilities, and profit bounds, we determine optimal pricing and delivery frequency, offering valuable insights into the feasibility and efficiency of the TC delivery model.

• 
  01:22 PM - 01:44 PM

  Coach Reservation for Groups Requests

  • Carlos Cardonha, presenter, University of Connecticut
  • Raghunathan Arvind, Mitsibushi Electric Research Laboratories

  Passenger transportation is a cornerstone of a railway company’s business, with ticket sales playing a pivotal role in revenue generation. This work extends traditional revenue management models by addressing the coach reservation problem for group requests, where two or more passengers must be seated in the same coach. We propose an exact mathematical programming formulation incorporating a first-come, first-served fairness condition for the offline case, in which all requests are known in advance. We also propose algorithms for online models of the problem. Our analysis for one of these models overcomes known barriers in the packing literature, yielding strong competitive ratio guarantees when group sizes are relatively small compared to coach capacity. Using data from the Shinkansen Tokyo-Shin-Osaka line, our numerical experiments demonstrate the practical effectiveness of the proposed policies.

• 
  01:44 PM - 02:06 PM

  Station-based Mobility-on-Demand Service Design

  • Shuiwang Chen, presenter, The Hong Kong Polytechnic University
  • Kai Wang, Tsinghua University
  • Lingxiao Wu, The Hong Kong Polytechnic University

  This paper addresses mobility-on-demand pricing problems which feature endogenous price-sensitive passenger choice (demand) and demand-responsive vehicle operations (supply). These problems involve a large-scale (mixed-integer) nonlinear nonconvex optimization structure, where passenger demand is governed by a discrete choice model, and vehicle operations is modeled using a closed queueing network model. We identify a hidden concave structure to circumvent the nonconvexity inherent in the choice model. Our approach involves developing a branch-and-cut methodology based on: (1) generalized Benders Decomposition cuts for the hidden concave structure; (2) convex relaxations to handle bilinear terms in queueing; and (3) a spatial branch-and-bound framework to explore global optimal regions and tighten relaxations. We apply this methodology to four mobility-on-demand pricing scenarios: urban aerial mobility, urban aerial mobility with service network design, one-way carsharing, and autonomous taxis. Extensive experiments demonstrate that the algorithm converges to global optimality and scales to large instances, outperforming state-of-the-art benchmarks.

• 
  02:06 PM - 02:28 PM

  Lagrangian Relaxation and Universal Fast Subgradient Method for Sea Freight Optimization

  • Aliaksandr Nekrashevich, presenter, Queen's University
  • Mikhail Kovalyov, National Academy of Sciences of Belarus
  • Genrikh Levin, National Academy of Sciences of Belarus
  • Yuri Levin, Queen's University
  • Guang Li, Queen's University
  • Mikhail Nediak, Queen's University

  We introduce a scalable stochastic approach for sea freight revenue management. Sea freight models are usually complex due to multi-faceted decisions across multiple business units, ranging from allotment negotiation and spot market dynamic pricing to container allocation and shipment scheduling. Optimization formulations are challenging to scale because of the network problem structure, driven by limited vessel capacity and port container balancing. We approach the model via Lagrangian Relaxation to relax the capacity constraints and substitute container management from the fluid approximation to decompose the original dynamic program into computationally tractable subproblems on the itinerary level. By applying the Universal Fast Subradient Method introduced by Yuri Nesterov over the Lagrangian multipliers, we achieve scalability due to the effective handling of the convex non-smooth structure of the total approximation objective.