We present a new system architecture for demand-side load management. The system is composed of modules for admission control, load balancing, and demand/response management that operate using online operation control, optimal scheduling, and dynamic pricing respectively. It can integrate different energy sources and handle autonomous systems with heterogeneous dynamics in multiple time-scales. Simulation results confirm the viability and efficiency of the proposed framework.
This is joint work with G. Costanzo, G. Zhu and G. Savard.
Effective management of cancer treatment facility for radiation therapy depends mainly on optimizing the utilization of linear accelerators. In this project, we are scheduling patients on those machines while taking into account their priority for treatment, the maximum waiting time before the first treatment and the duration of treatment. We collaborate with the “Centre Intégré de Cancérologie de Laval” to determine the best scheduling policy. The goal is to find, at each patient arrival, the best sequence of appointments that respects some constraints. Furthermore, we integrate the uncertainty related to the arrival of patients at the center. We develop an hybrid method combining stochastic optimization and online optimization to better meet the needs of central planning. Therefore, we use the information of future arrivals of patients to capture the most accurate picture of the expected utilization of resources. Randomly generated data allow us to study the behavior of our online algorithm according to its parameters. Tests on real data show that our method outperforms strategies typically used in such treatment centers.
One of the most powerful and simple approaches to model a customer’s choice behavior, with the aim to predict his choice decision facing different options, is non-parametric choice modeling of demand. In this approach, each arriving customer chooses from available alternatives according to an ordered preference list of products. If the customer's most preferred product is not available, he substitutes it with the next lower rank product in his ordered preference list.
In this paper, we propose a new mathematical programming approach to compute optimal allocation of airline resources under a non-parametric choice model of demand. We develop a modified column generation algorithm to efficiently solve large scale, real world practical problems. As the complexity of the algorithm increases with the number of the ordered preference lists, we provide an aggregation algorithm to reduce the number of the ordered preference lists without degrading the quality of the solution. The computational results show that the approach outperforms alternative models.
We developed optimization models for many personnel or vehicle scheduling problems for airline, bus and rail transportation. Each model includes thousands of constraints and millions of millions of variables. We constructed new mathematic methods solving these large problems by working only on a subset of variables and constraints at the time. The dynamic adjustment of the subsets permits to catch the pertinent information and achieve the optimal solution. These systems are commercialized around the world by AD OPT and Giro, two spin-off from the university who hire more than 400 scientists.