3 présentations

• 
  08h30 - 08h52

  Distributionally Robust Group Testing with Correlation Information

  • Zhuoyu Long, Chinese University of Hong Kong
  • Jin Qi, Hong Kong University of Science and Technology
  • Yu Sun, Chinese University of Hong Kong
  • Aiqi Zhang, prés., Wilfrid Laurier University

  Motivated by the need for more efficient and reliable methods of group testing during widespread infectious outbreaks, this paper introduces a novel operational improvement to the classic group testing procedure, in which a single test is conducted on the pooled sample, followed by individual testing of positive pools. Our method minimizes a weighted sum of tests and misclassifications, predicated on the known prevalence rates and inter-individual Pearson correlation coefficients. Recognizing the inherent ambiguity in the population-level probability of infections due to these correlations, our approach leverages a distributionally robust optimization (DRO) framework to counteract the worst-case probability distribution. In single-cluster cases, where each pair of subjects are equally correlated, we establish the optimality of the widely adopted uniform group sizes. We further show that larger testing groups are generally favored under high correlation, whereas individual testing becomes optimal under high prevalence. In multi-cluster cases, where the population is formed by several intra-correlated but inter-independent clusters, we highlight the effectiveness of mixed-cluster testing strategies, particularly at lower levels of prevalence and correlation. For both single- and multi-cluster cases, we develop polynomial-time solutions and investigate the change of optimal pooling strategy as a function of imperfect tests. We demonstrate the benefits of adopting the DRO framework through a comprehensive comparison with stochastic alternatives, and we illustrate the significant impact of considering correlated infections through a case study on a real-world dataset.

• 
  08h52 - 09h14

  Conformal Inverse Optimization for Adherence-aware Prescriptive Analytics

  • Bo Lin, prés., University of Toronto
  • Timothy Chan, University of Toronto
  • Erick Delage, HEC Montréal

  Inverse optimization is increasingly used to estimate unknown parameters in an optimization model based on decision data. We show that such a point estimate alone is insufficient in a prescriptive setting where the estimated parameters are used to prescribe new decisions. The resulting decisions may be low-quality and misaligned with human intuition and thus are unlikely to be adopted. To tackle this challenge, we propose a novel decision recommendation pipeline, which seeks to learn an uncertainty set for the unknown parameters and then solve a robust optimization model to prescribe new decisions. We show that the suggested decisions can achieve bounded optimality gaps, as evaluated using both the ground-truth parameters and human perceptions. Our method demonstrates strong empirical performance compared to the standard inverse optimization pipeline. Finally, we perform a case study where we apply this new pipeline to provide delivery route recommendations in Toronto, Canada. Our approach achieves a significantly higher delivery path adherence rate than current industry practices without compromising service quality. Moreover, our method provides a better trade-off between absolute and perceived decision quality than baselines under various realistic scenarios, including cases with model mis-specification and data scarcity.

• 
  09h14 - 09h36

  Saving Whales with Optimal Control

  • Yu Gong, prés., Smith School of Business, Queen's University
  • Jue Wang, Smith School of Business, Queen's University
  • Jingwei Zhang, SC Johnson College of Business, Cornell University

  Collision with large vessel is the leading cause of whale mortality. The collision risk can be reduced by the establishment of vessel speed reduction (VSR) zone, in which large vessels are asked to reduce speed. However, reducing speed is costly to the mariners and causing shipping delays, which leads to a low compliance rate. We analyze the satellite tracking data to understand how mariners dynamically adjust their speed near VSR, and use this knowledge to improve the VSR design. The goal is to jointly minimize the whale mortality and the economic impact to the shipping industry.