4 présentations

• 
  10h15 - 10h40

  Opmitizating Rolling Stock Maintenance Planning for a French Public Transport Operator

  • Louis-Pierre Campeau, prés., Artelys
  • Carlyle Deligny, Artelys

  Maintenance strategy is crucial to minimize operating costs, optimize the operational use of resources, and ensure that the organization’s contractual commitments to its subcontractors and employees are respected. This presentation will introduce a business case Artelys led in the field of transportation. Our solution imports data from the customer’s Computerized Maintenance Management System (CMMS) and optimizes the maintenance strategy on a multi-horizon basis: strategic, tactical and pre-operational. The first part of the project consisted in long-term strategic Sales and Operations Planning (SOP) generation as well as medium-term tactical Master Production Schedule (MPS) generation and follow-up. The software was embedded in the company existing system, modeling thousands of components and constraints. After this successful implementation, Artelys is working on an upgraded optimized pre-operational planning that will further enhance the decision-making performance and understandability. The presentation will give a detailed overview of the company-wide solution Artelys developed based on Artelys Crystal Resource Optimizer, where the typical size of the problem is 100 users, 1000 trains, 30 000 consumable spare parts and 50 000 work orders.

• 
  10h40 - 11h05

  Anomaly Detection for Refrigeration Ice-Up

  • Jean-François Landry, prés., Ivado Labs
  • Louis-Philippe Bigras, Ivado Labs
  • Steve Liu, Ivado Labs
  • Pierre-Luc Bacon, Ivado Labs
  • Louis-Martin Rousseau, Polytechnique Montréal
  • Thibaut Vidal, CIRRELT, Université de Montréal & ICD-LOSI, Université de Technologie de Troyes

  In commercial refrigeration, ice accumulation on evaporator coils—commonly referred to as “case ice-up”—can lead to energy inefficiencies, product loss, and costly downtime due to the need for manual defrosting. This presentation explores a practical use case aimed at early detection and mitigation of case ice-up events in refrigerated display cases. By leveraging temperature sensor data, particularly during defrost cycles, the system identifies deviations from expected thermal patterns. Specifically, it monitors the pull-down phase following defrost: a healthy case shows a sharp temperature drop, while a gradually declining temperature may signal ice buildup. The proposed solution automates detection of this abnormal behavior, extends the duration of subsequent defrost cycles as a corrective action, and escalates to an alarm if the issue persists. This tiered response minimizes disruption, improves energy efficiency, and helps identify other potential case health issues—such as faulty fans, broken doors, or high ambient humidity. The result is a smarter, more proactive approach to maintaining refrigeration performance and operational reliability in retail environments.

• 
  11h05 - 11h30

  AI-Driven Revenue Growth Management and Emerging Generative AI Applications at ALDO Group

  • Fatih Nayebi, prés., ALDO Group
  • Adrien Rimélé, Ivado Labs
  • Adrian Alarcon Delgado, ALDO Group
  • Rodrigo Alves Randel, Ivado Labs
  • Dorian Dumez, Ivado Labs
  • Onur Erkin Sucu, ALDO Group
  • Arnav Gupta, ALDO Group
  • Clara Lacroce, Ivado Labs
  • Qing Liu, Ivado Labs
  • Bani Mehri, ALDO Group
  • Carl Perreault-Lafleur, Ivado Labs
  • Claudio Sole, Ivado Labs
  • Mehdi Towhidi, Ivado Labs
  • Soraya Yama, Ivado Labs
  • Yossiri Adulyasak, HEC Montréal
  • Maxime Cohen, McGill university
  • Jean-François Cordeau, HEC Montréal, GERAD, CIRRELT
  • Andrea Lodi, CERC, Polytechnique Montréal, Montréal, Canada and Jacobs Technion-Cornell Institute, Cornell Tech and Technion - IIT, New York, USA

  This talk, co-presented by ALDO Group and IVADO Labs, showcases a collaborative AI initiative focused on Revenue Growth Management within the retail sector. ALDO, a global fashion retailer, has partnered with IVADO Labs to develop AI-powered solutions that integrate predictive modeling and mathematical optimization across three key domains: demand forecasting, markdown optimization, and order fulfillment. These efforts aim to enhance forecast accuracy, optimize pricing decisions, and streamline supply chain operations, resulting in improved inventory turnover, reduced operational costs, and increased customer satisfaction.
  In parallel, ALDO Group has been exploring Generative and Agentic AI technologies to drive innovation in customer experience and internal operations. Initiatives include a generative AI-powered employee chatbot, an agentic AI call center prototype, automated generation of product and SEO descriptions, and an annual Retail GenAI Hackathon co-hosted with McGill University.
  The presentation will highlight technical approaches, real-world deployment challenges, and the value of industry-academic collaboration in accelerating applied AI innovation in retail.

• 
  11h30 - 11h55

  Integrated Approach for Maintenance Optimization: Presentation of an Operational Business Case in Transportation

  • Louis-Pierre Campeau, prés., Artelys

  Maintenance strategy is crucial to minimize operating costs, optimize the operational use of resources, and ensure that the organization’s contractual commitments to its subcontractors and employees are respected.
  This presentation will introduce a business case Artelys led in the field of transportation. Our solution imports data from the customer’s Computerized Maintenance Management System (CMMS) and optimizes the maintenance strategy on a multi-horizon basis: strategic, tactical and pre-operational. The first part of the project consisted in long-term strategic Sales and Operations Planning (SOP) generation as well as medium-term tactical Master Production Schedule (MPS) generation and follow-up. The software was embedded in the company existing system, modeling thousands of components and constraints. After this successful implementation, Artelys is working on an upgraded optimized pre-operational planning that will further enhance the decision-making performance and understandability.
  The presentation will give a detailed overview of the company-wide solution Artelys developed based on Artelys Crystal Resource Optimizer, where the typical size of the problem is 100 users, 1000 trains, 30 000 consumable spare parts and 50 000 work orders.