CORS / Optimization Days
HEC Montréal, May 29-31, 2023
CORS-JOPT2023
HEC Montreal, 29 — 31 May 2023
CORS Practice prize
May 30, 2023 03:30 PM – 05:10 PM
Location: CPA du Québec (green)
3 Presentations
-
03:30 PM - 03:55 PM
Weather Routing for Maritime Vessels
True North Marine (TNM) is a Montreal-based consulting firm offering services on route selection and analysis to ocean-going bulk vessels. Each day, hundreds of vessels are supported by some 40 route planners in selecting the best route and speed giving the current weather forecast. Automatic weather routing is very complex. The detailed weather information is available as forecasts in speed and direction for wind, waves, and currents for every three hours in the next seven days. Each route requires its own large network resulting in challenging optimization. We describe a system with many analytics modules developed that shows large potential savings.
-
03:55 PM - 04:20 PM
Optimizing milk pools at the Rogers Hixon Ontario Human Milk Bank
Human donor milk is vital for millions of preterm infants born each year, but the macronutrient content of donations can vary significantly. In collaboration with the Rogers Hixon Ontario Human Milk Bank, we developed a data-driven framework to pool multiple donations using machine learning and optimization. Over a one-year trial, our implementation yielded significantly higher macronutrient content than current pooling practices, with the proportion of pools meeting clinical fat and protein targets increasing by approximately 31%, with a 60% decrease in recipe creation time.
-
04:20 PM - 04:45 PM
Real-time management of traffic lights – improving mobility and decreasing greenhouse gas emissions
Congestion in urban centers causes stress, delays, increased fuel consumption and emissions. In dense areas, many consecutive traffic lights impose significant stop-and-go for all vehicles, particularly harmful for trucks and buses. A single stop avoided for a heavy truck can represent up to 2.5 liters of fuel saved; they accelerate slowly, are noisy, and cause further delays for the vehicles behind them. Moreover, helping buses avoid red lights supports adherence to the schedule. Avoiding stops at red lights for these vehicles is a win-win situation for their operators and for all the remaining traffic in the city. This project consists of a successful collaboration between our research team, the city of Trois-Rivières, the Société de Transports de Trois-Rivières, and industrial partners. We focused on 14 traffic lights on an important avenue in the city. Our collaboration enabled these lights to be connected to an intelligent traffic management system to grant priority to trucks and buses, in real-time and without human intervention. By avoiding stops for these large and heavy vehicles, we obtain multiple benefits in terms of time, efficiency, economy, and emissions. We performed a detailed modeling of the network and traffic of Trois-Rivières. We developed a simulation tool with the latest origin-destination survey from the Ministère des transports du Québec, validated manual and automated counters, and calibrated the results with observations from LiDAR scanners installed at strategic points in the city. Our microsimulation models a large territory of 39 km2 , for a total of 659.85 km of road network comprising over 5 480 edges, handling 289 000 vehicles per day, which is significantly larger than most simulation studies. We implemented the traffic signal phasings used by the city and developed optimized signal priority rules that were systematically tested with the simulator. Our results indicate that a parsimonious prioritization strategy has huge benefits for all parties. The strategy delays the green light at the end of a cycle for no more than 15 seconds, or truncates the red phase at most 15 seconds in advance, to grant priority for an incoming truck or bus that would otherwise have stopped. We can avoid 8% of stops for the trucks, improving their travel time and decreasing fuel consumption by 1.9%, which corresponds to 10 tonnes of GHG emitted over the course of a year for a single truck. The impact on the overall traffic is negligible, and often positive: during the PM peak traffic, granting priority for the trucks also improves fluidity for the overall traffic, improving their travel times and decreasing overall fuel consumption by 0.11%. For the transit provider, we observed up to 846 occasions of buses of a given line being more than 1 minute late in the simulation, and granting them priority reduces it by up to 47%. The average delay over the line can be virtually removed. The impact on the surrounding vehicles is shown to be negligible or even slightly positive: average travel times and waiting times for society can be slightly improved, with decreased consumption and emissions. These results are currently being implemented by the city administration with the support of a technology partner while the city updates its traffic signal hardware. A private fleet of trucks has been equipped with real-time GPS monitors and the results of this pilot study systematically confirm the simulated benefits. Buses are also currently being granted priority based on the policies we developed, and the results are also being confirmed in reality. This project has demonstrated that advanced analytics, optimization, and a strong collaboration between industry and academia can provide strong results for society. Benefits include decreased travel times, improved transit systems, better financial performances, and lower emissions.