In addition to retailing new vehicles, automotive manufacturers in the United States sell millions of vehicles through leasing and to fleet customers every year. The majority of these vehicles are returned to the automotive manufacturers at the end of the contracted term and must be “remarketed.” In 2007, about 10 million used vehicles were sold at more than 400 auctions in the United States. Large consigners face decisions every day about when, where, and at what price to offer these vehicles, which has significant financial implications for their profitability.
To address the challenges of the distribution process, Power Information Network (PIN), a division of J.D. Power and Associates, developed the PIN Optimal Distribution of Auction Vehicles System (ODAV), an automated decision optimization system that helps remarketers maximize profits through the most advantageous distribution of their auction vehicles. At the core of the system is a combination of three models that determine the distribution of the vehicles on a daily basis: a nearest neighbor linear regression model for short-term auction price forecasting; an autoregressive integrated moving average time-series analysis model for volume-price elasticity; and a genetic algorithm optimizer for vehicle distribution.
Since its launch in 2003, PIN has been providing ODAV services on a daily basis, and to date, more than two million vehicles have been distributed through this system. In this paper, we will describe the PIN ODAV System, its implementation, and the business impact by using as an example the experience with our largest client, Chrysler Group LLC.
[1]
Goldberg,et al.
Genetic algorithms
,
1993,
Robust Control Systems with Genetic Algorithms.
[2]
Zbigniew Michalewicz,et al.
Genetic Algorithms + Data Structures = Evolution Programs
,
1996,
Springer Berlin Heidelberg.
[3]
Gwilym M. Jenkins,et al.
Time series analysis, forecasting and control
,
1972
.
[4]
Giorgio Gambosi,et al.
Complexity and approximation: combinatorial optimization problems and their approximability properties
,
1999
.
[5]
Robert E. Hall,et al.
Economics: Principles and Applications
,
1997
.
[6]
David E. Goldberg,et al.
Genetic Algorithms in Search Optimization and Machine Learning
,
1988
.