Showing papers by "André Langevin published in 2000"

Value Analysis and Optimization of Reusable Containers at Canada Post

Abstract: We analyzed the use and sharing of a logistics package, a returnable container, often out of stock, between Canada Post and its large mailing customers. Standardizing and sharing logistical packaging should bring members of a supply chain tangible benefits from productivity and efficiency gains. We determined that inventory imbalance between supply locations and demand locations caused the problem. We found that different accounting methods gave very different results in calculating the benefits. We concluded that there is always a benefit for the mailer. For Canada Post to benefit, however, the container must be returned quickly. We used a minimum-cost-flow model to confirm that the company has enough containers to satisfy demand. With better planning and control of the inventory, Canada Post can prevent disequilibrium between supply and demand.

Roadway Snow and Ice Control

Abstract: Roadway snow and ice control (RSIC) is one of the most complex and fascinating venues for arc routing applications. Arc routing problems occur in several different aspects of RSIC, including plowing roadways and sidewalks, spreading chemicals (salt and other de-icers) and abrasives (sand and stone), and in using snowblowers to load snow into trucks for hauling to disposal sites. The limited progress of computerized routing packages in this area (Office of the Legislative Auditor, 1995) highlights the difficulty of the problems and the weaknesses of theoretical arc routing models for snow and ice control. Part of the difficulty is due to the complex operational and infrastructure constraints, especially in urban areas. The dynamic nature of the problems also adds complexity, as conditions can vary dramatically over time and space. Furthermore, a wide variety of equipment with different operating characteristics can be deployed for RSIC.

A Hybrid Travel Distance Approximation for a GIS-based Decision Support System

Abstract: Logistics and transportation decision support systems rely on models of travel distance for evaluating scenarios and estimating costs. This paper presents a new method to model travel distances that combines the accuracy of shortest path travel with the simplicity of a distance approximation. The distance approximation was developed and tested in a geographic information system (GIS)-based decision support system for snow removal and disposal in Montreal, Canada.

Transformations and Exact Node Routing Solutions by Column Generation

Abstract: This chapter examines arc routing problems from the “dual” perspective of node routing. It first attempts to explain and respond to a question from a “typical” engineering and applied mathematics graduate who was exposed to classical node routing problems such as the transportation and the traveling salesman problems, and ventured a little into the vast literature on problems and solutions for the different variants of vehicle routing (node routing) problems. Why shouldn’t (or why should) any arc routing problem be viewed as some version of a node routing problem, pending an appropriate transformation of the corresponding graph? The first part of this chapter will examine this question addressing the issue of when such a transformation is necessary and the complementary question of when, or for what arc routing problems, from computational point of view, a transformation to node routing is inappropriate. In addition, the first part will attempt to provide a partial account of the different transformation schemes proposed over the years for arc routing problems into node routing setting. For an excellent write-up of exact solution methodologies for “hard” arc routing problems addressed without transformation to a node routing setting the reader is directed to chapters in this book by Eglese and Letchford, Benavent, Corberan, and Sanchis, and Johnson.