Fleet Management Planning

Case Study used in Ming-Che Lai, Han-Suk Sohn, Dennis L. Bricker, “A
hybrid Benders/genetic algorithm for vehicle routing and scheduling problem”
submitted for publication in International Journal of Industrial Engineering –
Theory, Applications, and Practice.

In this research, the hybrid Benders/genetic
algorithm was implemented to the fleet management planning of the Rockwell
Collins, Inc. headquartered in Cedar Rapids, Iowa. We provide here some extra information for this case
study as below:

Figure 1. Schematic Map of Travel Times

In this case study, travel times between the stops
were measured and shown on the schematic map in Figure 1. Now, for the purpose
of implementation, each building was labeled from 1 to 25 and the correspondences
between stop number and building are shown in Table 1. Since the travel times
between each stop can be defined to be the distances of the corresponding link,
we compute the shortest paths between every pair of nodes and record them in
the travel time matrix of a complete network (see Figure 2). These travel
times, of course, could be increased under adverse traffic and weather
conditions.

Table 1. Correspondences of
Stops and Buildings

Table 2. Given Length for
Each Links

Figure 2. Shortest Path Matrix

The best solution found from
the hybrid algorithm has 137 minutes of total travel time, which consists of 48
minutes, 46 minutes, and 43 minutes for each route (see Figure 3). The travel
times, as expected, are distributed evenly among three routes in this result. Some
stops are visited each time the route is traversed, while others are visited
less frequently, according to the volume of mail and dispatches generated
there.

Figure 3. Proposed Route Plan

Computer programming codes
are available upon request. Please email
to Professor Sohn for
further enquiries regarding this case study.