Page 44 - DCAP208_Management Support Systems

Unit 3: Decision Support Systems

Notes
Problem Overview
We developed the system for a US-based car manufacturer that has more than 1 million
cars returned from leases or rentals each year. The manufacturer owns the cars, and the
problem is how to best distribute these cars among hundreds of auction sites around the
United States. The cars vary by make and model, mileage, options, wear and tear, and so
on. These characteristics, along with others, influence the car’s sale price at each particular
auction. Our central challenge was to achieve the “best” possible distribution among
these auction sites—that is, the distribution that maximizes the net proceeds from all
sales. The process of making optimal recommendations involves many considerations,
ranging from price prediction for various car types at different locations, to price
depreciation and volume effects, to transportation issues. One million cars per year
corresponds to approximately 4,000 cars per working day. So, each day, a remarketing
team must make 4,000 decisions regarding which auction site will maximize the sales
price of each car. Further, due to volume effects, assigning cars to auctions is highly
interrelated, and therefore it’s not possible to process these cars sequentially.
Figure 1: Adaptive Business Intelligence

Say, for example, that a company uses 50 auction sites and processes a mere 1,000 cars per
day. This results in a mind-boggling 501,000 distribution choices! No computer can check
out all these possible combinations in a human lifetime. Nevertheless, the manufacturer
requires decisions on all of the cars today.

Problem Complexity
To illustrate the task, we’ll use a silver, four door 2002 Toyota Corolla with 34,983 miles,
a sun roof, automatic transmission, power windows, power seats, and many other options.
At the moment, the car sits at a dealership in Virginia, and we must decide where to send
it. At first glace, this looks easy. We might be tempted to simply look up the car’s average
sales price at each auction using one of many guides, such as the Black Book, Kelley Blue
Book, or Manheim Auction Report. After adjusting the price based on the car’s mileage,
options, and so on, and estimating transportation costs—both manageable calculations—
we might simply decide to target the auction with the highest current average sales price.
So, what’s the problem? In a word: volume. Per car, it’s cheaper to ship a truckload of cars
from one place to another than it is to ship one or a few cars at a time.

For more than 14 cars, we usually calculate transportation costs by determining the fee for
transporting 14 cars (or multiples of 14 cars), then calculate the remainder at the applicable
rate. For example, transporting 20 cars would cost us $85 per car for the first 14 cars, and
then $120 per car for the remaining 6 cars, for a total cost of $1,910. Beyond transportation

Contd....

LOVELY PROFESSIONAL UNIVERSITY 37