Search

In this paper, we introduce the Special Issue on Machine Learning Methods and Applications in Large-Scale Route Planning Problems, which draws its inspiration from the academic community's positive reception of the 2021 Amazon Last Mile Routing Research Challenge. We provide a structured overview of the papers featured in this special issue, and briefly discuss other noteworthy contributions to the research challenge. Further, we point the reader to a number of peer-reviewed publications outside of this special issue that directly or indirectly emerged from the research challenge. We conclude by highlighting a number of important priorities for future research into applications of machine learning to real-world route planning problems. [ABSTRACT FROM AUTHOR]

The 2021 Amazon Last Mile Routing Research Challenge, hosted by Amazon.com's Last Mile Research team, and scientifically supported by the Massachusetts Institute of Technology's Center for Transportation and Logistics, prompted participants to leverage real operational data to find new and better ways to solve a real-world routing problem. In this article, we describe the data set released for the research challenge, which includes route-, stop-, and package-level features for 9,184 historical routes performed by Amazon drivers in 2018 in five metropolitan areas in the United States. This real-world data set excludes any personally identifiable information: all route and package identifiers have been randomly regenerated and related location data have been obfuscated to ensure anonymity. Although multiple synthetic benchmark data sets are available in the literature, the data set of the 2021 Amazon Last Mile Routing Research Challenge is the first large and publicly available data set to include instances based on real-world operational routing data. History: This paper has been accepted for the Transportation Science Special Issue on Machine Learning Methods and Applications in Large-Scale Route Planning Problems. [ABSTRACT FROM AUTHOR]

Daily tactical planning of the vehicle fleet is critical for a car rental business. In this decision-making process planners have to address the conflicting concerns of field management, namely, assuring adequate vehicle availability to satisfy customer demand while maintaining a high degree of utilization for each car in the fleet; this is achieved by the appropriate fleet deployment among car rental locations. This work was inspired by a series of problems encountered by the decision-makers of a major US car rental company. We first present the formulation of the tactical fleet planning model. Based on its structure, we decompose the model into a fleet deployment model and a transportation model. We develop optimal conditions for both subproblems and then we present a heuristic to reduce the gap from optimality for the complete model. We further present the application of the entire solution methodology via a case analysis for the State of Florida. Finally, we develop three extensions to the fleet deployment model to capture the cost of unsatisfied demand and fleet surplus, the service level, and a general price-demand function. We provide optimal closed-form conditions for all the extensions., 1. Introduction One of the primary functions of revenue management for a car rental company is to determine the optimal mixture and size of the vehicle fleet that should be [...]

The Crew ResourceSolver decision-support system employs advanced optimization modeling and solution techniques to solve complex, large-scale pilot staffing and training problems at Continental Airlines. The system determines optimal pilot transitions and efficiently allocates and schedules training resources. The results are improved staffing levels and substantial cost savings through reduced staffing, hiring, and training costs. Continental has estimated savings of over $10 million per year using the system. [ABSTRACT FROM AUTHOR]