Applications of OR VI

Contributed Session

Time Slot: Friday Afternoon

Room: AUD_B

Chair: Carlo Musmanno

Machine Learning heuristic for Variable Cost and Size Bin Packing Problem with Stochastic Items

Time: 15:40

Stanislav Fedorov (DAUIN & CARS@Polito Politecnico di Torino, Torino, Italy), Fadda Edoardo, Perboli Guido

Third-party logistics becomes an essential component of efficient delivery, enabling companies to rent transportation services instead of keeping an expensive fleet of vehicles. However, the contracts with the carriers usually have to be booked beforehand when the delivery demand is unknown. This decision process is strongly affected by uncertainty, provided with a long (tactical) planning horizon, and can be expressed as choosing an appropriate set of bins (fleet contracts). Formally, it can be modeled as the Variable Cost and Size Bin Packing Problem with Stochastic Items [1]. It consists of packing the set of items (goods) with uncertain volumes and quantities into containers (bins) of different fixed costs and capacities. This problem is described via a two-stage stochastic programming approach, where the cost of the bins of the second stage is significantly higher. Since it cannot be solved for large realistic instances by means of exact solvers for a reasonable time and memory consumption, this paper introduces a Machine Learning heuristic to approximate the first stage decision variables. Several numerical experiments are outlined to show the effectiveness of the proposed approach to deal with realistic instances of up to 3000 items. Further, the proposed heuristic is compared to the recent Progressive Hedging-based heuristic and showed a significant computational time reduction. Finally, different classification approaches are compared, and the feature selection process is explained to gain insight into heuristic performance to deal with the outlined problem.

[1] Crainic, T. G., Gobbato, L., Perboli, G., Rei, W., Watson, J. P., & Woodruff, D. L. (2014). Bin packing problems with uncertainty on item characteristics: An application to capacity planning in logistics. Procedia-Social and Behavioral Sciences, 111, 654-662.

Last-mile delivery with vans and autonomous robots: an analysis on the impact of the return policy

Time: 16:00

Mikele Gajda (HEC Lausanne (Faculty of Business and Economics), University of Lausanne), Boysen Nils, Gallay Olivier

In recent years, technological advancements have shaped innovative and promising last-mile delivery solutions. One of these approaches involves the employment of autonomous delivery robots (ADRs) in conjunction with vans from which the robots can be picked up and released.These robots operate at pedestrian pace on the sidewalk, providing a reliable operator for the last delivery step. In our research, we study a multi-modal transportation system in which vans convey ADRs from depots to prospective drop-off places, with the final portion of the delivery being handled autonomously by the robots, which then return to one of the available depots. The goal of this contribution is to investigate and compare different policy approaches for the return of the robots to the depots. Indeed, the choice of the return policy has a direct impact on the number of robots required to perform a given delivery plan. Our main objective is to minimize the number of used robots to complete a certain amount of delivery jobs. We define and analyze three possible robot return policies in detail (dedicated-station policy, closest-station policy, most-suitable-station policy), and we propose suitable optimization algorithms for each of them in order to find optimal solutions in polynomial time. Our findings, obtained from computational experiments performed on an extensive set of realistic instances, demonstrate that the robot return policy should be carefully chosen in order to achieve the targeted delivery plan with a minimum number of robots.

Selection of Cultural Sites via Optimization

Time: 16:20

Roberto Musmanno (Department of Mechanical, Energy and Management Engineering), Annarita De Maio, Aurora Skrame, Francesca Vocaturo

We focus on a combinatorial optimization problem arising in regional tourism management. The problem consists in selecting cultural sites to be promoted among others distributed in a regional geographic area. Specifically, when a cultural site is selected, the potential visitors who decide to travel by train have the opportunity to take advantage of a free-of-charge shuttle dispatched from the closest railway station to the site. For the solution of this problem we propose an optimization framework based on the formulation of a bi-objective mathematical model. The computational results are presented by considering a case study derived from a regional project.

Evaluation of economic impacts for a demand response program within a sawmill

Time: 16:40

Imen Chaabouni (Laval University, Quebec), Lehoux Nadia

In this paper, we propose an economic evaluation method for a typical sawmill in Quebec, Canada, which is considering participating in a Demand Response (DR) program such as that provided by Hydro-Quebec, the local electricity utility. In this method, the variation in costs associated with program participation is tracked, including unmet demand, inventory utilization, and electricity bill costs against potential revenues generated from participation in DR events. These are evaluated for different scenarios with a varying number of DR hours that could occur during winter. For this purpose, a suitable mathematical model was selected from the literature. Using real data collected between December 2019 and February 2020 and employing techniques based on demand shedding and demand shifting at the planing stage, the results revealed that participating in Hydro-Quebec’s DR program is possible with limited backorder quantities. It was also shown that each additional production hour added to the weekly calendar, load shifting strategy, reduced costs by about 70%, in the case of 80 DR hours per winter season. Furthermore, a 0.44% cost reduction can be achieved for DR events of 16 hours by adding 2.5 hours per week of overtime to the production calendar (about C\$ 17,665.40 profit).