The following zip-files:
all contains 5 text files with data instances used in the paper below to test different heuristics for the multiple-depot vehicle scheduling problem.
A.-S. Pepin, G. Desaulniers, A. Hertz and D. Huisman, "Comparison of heuristic approaches for the multiple depot vehicle scheduling problem", Report EI2006-34, Econometric Institute, Erasmus University Rotterdam (2006), 23 pages.
Download: pdf file.
All instances are generated in the same way as in Capenato et al. (1989). The name of the file indicates the number of depots ("m"), the number of trips ("n") and finally the number of the instance (0,1,2,3,4). The format is as follows:
(first line) m <tab> n <tab> for each depot the maximum number
of vehicles
(second line and further) cost matrix of size (m + n) x (m + n).
The number in row i, column j indicates
- if i <= m and j > m: the cost of pull-out trip from depot i
to trip j-m (including 50% of fixed cost for the vehicle),
- if i > m and j <= m: the cost of pull-in trip from trip i-m to depot j
(including 50% of fixed cost for the vehicle),
- if i > m and j > m: the cost of performing trip j-m after trip i-m.
A -1 indicates that it is infeasible to perform trip j after trip i.
Note that, i <= m and j <=m has no meaning and there are all -1's
there.
The best available solutions are reported in the table below. These solutions are either the optimal solutions obtained by solving with CPLEX (indicated with a *) or the best solution from the different heuristic approaches described in the paper above (a = truncated branch-and-cut, b = truncated column generation, the other heuristics never provided the best solution).
| 4 depots | 8 depots | |||||
| instance | 500 tasks | 1000 tasks | 1500 tasks | 500 tasks | 1000 tasks | 1500 tasks |
| s0 | 1289114 * | 2516247 a | 3830912 a | 1292411 b | 2422112 a | 3500160 b |
| s1 | 1241618 * | 2413393 b | 3559176 b | 1276919 * | 2524293 a | 3802650 b |
| s2 | 1283811 * | 2452905 a | 3649757 a | 1304251 * | 2556313 b | 3605094 b |
| s3 | 1258634 * | 2490812 a | 3406815 a | 1277838 b | 2478393 a | 3515802 b |
| s4 | 1317077 * | 2519191 * | 3567122 a | 1276010 * | 2498388 a | 3704953 b |
In the following table, the optimal number of vehicles is given.
| 4 depots | 8 depots | |||||
| instance | 500 tasks | 1000 tasks | 1500 tasks | 500 tasks | 1000 tasks | 1500 tasks |
| s0 | 123 | 241 | 368 | 124 | 232 | 337 |
| s1 | 118 | 229 | 338 | 123 | 244 | 366 |
| s2 | 123 | 233 | 350 | 126 | 247 | 349 |
| s3 | 120 | 237 | 326 | 123 | 237 | 338 |
| s4 | 126 | 238 | 343 | 123 | 240 | 360 |
This zip-file contains 140 files with the data of
random data instances described in the following paper:
D. Huisman, R. Freling and A.P.M. Wagelmans, "Multiple-Depot Integrated Vehicle and Crew Scheduling", Report EI2003-02, Econometric Institute, Erasmus University Rotterdam (2003), 26 pages.
Download: pdf-file.
Mind out! There is a small error in Table 1: the end time of type 1 ("early duty") should be 16:30 instead of 15:30.
The files are named as follows:
If you have any questions, do not hesitate to contact me.