forked from geoffleyland/VRPH
-
Notifications
You must be signed in to change notification settings - Fork 1
/
README
executable file
·95 lines (76 loc) · 3.77 KB
/
README
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
This version of VRPH is a fork of a fork of [COIN's VRPH]
(https://projects.coin-or.org/VRPH), with a number of bugs
fixed, features added, and code and usability improvements.
The first commit is a copy of VRPH 1.0, and the major
changes to it are:
- Multiple day cases are read correctly
(geoffleyland, 83d1433)
- Performance improvements for Flip operation
(geoffleyland, 4325f07, d0e9479)
- Many complier warnings and non-critical bugs fixed,
code cleanup (geoffleyland, juherask)
- Improved and cleaner output
(geoffleyland, juherask)
- Command line use of different applications is unified
(juherask, 573cc5a)
- All applications can be given a cutoff parameter that
interrupts optimization after specified time.
(juherask, 573cc5a)
- One can specify seed for RNG, and all streams are
initialized (juherask, 77edebc)
- Several critical bugs in 3-opt move has been fixed
(juherask, 48b1fd6)
Additionally, there is a branch local_search_stats which
includes additional data gathering on local search operations.
What follows is the original README:
VRPH is a library of heuristics for quickly generating
solutions to Vehicle Routing Problems. VRPH was
developed by Chris Groer as part of his dissertation at
the University of Maryland, College Park. This code is
provided in hopes that it will be useful to others researching
computational methods for the VRP.
To compile VRPH, read the INSTALL file. There are no
external dependencies and it should be straightforward
to build the code. Building the code produces at least
4 binaries, all located in the VRPH-1.0/bin directory:
1) vrp_rtr - implementation of a metaheuristic for quickly
generating good solutions to the VRP
2) vrp_sa - implementation of a simulated annealing
algorithm for the VRP. In general, solutions are not as
good as those produced by vrp_rtr.
3) vrp_ej - implementation of a neighborhood ejection-
injection algorithm for the VRP. This can produce solutions
that are better than vrp_rtr and can be used to further
refine these solutions
4) vrp_init - implementation of two classical heuristics
for generating initial feasible solutions
There are several additional binaries that require you to
link with external software:
1) vrp_sp - implementation of a set partitioning-based
algorithm for the VRP. Many feasible routes are generated
using the metaheuristic in vrp_rtr. These are added as
columns/variables to the set partitioning problem.
This requires you to link with COIN-OSI and GLPK. It should
not be hard to use a different MIP solver as there is very
little solver-specific code
2) vrp_plotter - a utility to write graphical illustrations
of solutions to a .ps file. This requires one to link to
the open source PLPlot library.
The VRPH-1.0/doc directory is a placeholder for Doxygen-generated
documentation. See the INSTALL file for details on creating
this documentation via the makefile.
Right now, the code is able to generate solutions to classical
capacitated VRPs. Adding support for time windows, multiple
depots, etc. should not be too difficult and the code was
designed to make it somewhat easy to add new constraints.
The various subdirectories contain README's that describe
the data sets, etc. See the LICENSE file in this directory for
details on the licensing of this code.
The run_all_tests.sh script runs the vrp_rtr binary on all of the
data files. The compare_apps.sh script allow one to compare the
performance of the different applications by running them with
a fixed set of parameters on different sets of benchmark problems.
If you find the code useful, I'd like to know, so send me an email.
If you encounter bugs or problems, I'd like to know that too.
Chris Groer