File Exchange

image thumbnail

Cross Entropy TSP Solver

version 1.0.0.0 (68.7 KB) by Sebastien PARIS
Solve TSP problem with a CE method.

2 Downloads

Updated 23 Jul 2008

View License

TSP toy problem solved by Cross-Entropy Method (stochastic optimization procedure) based on generation of multiple random paths.

Comments and Ratings (12)

Tammy Zeng

why it doesn't work? It says some functions lack one input

liudaohai liudaohai

Thank you very much!!

Sébastien Paris

Update with missing file (cost_tsp) coming soon. Thx for reporting this.

liudaohai liudaohai

??? Undefined command/function 'cost_TSP'.

Error in ==> TSP_CE at 155
[S , indice] = sort(cost_TSP(L , X));

Error in ==> demo_TSP_CE at 33
[X_opt , S_opt , Pt , T] = TSP_CE(L , X1 , option);

liudaohai liudaohai

file demo_TSP_CE does not work,after recompile both mex files

Zdravko B.

very useful and well written. Thanks for sharing

Sébastien Paris

You have to recompile both mex files if your are not using a win32 system.

Mats Ahlgren

does not work; even example in "help TSP_CE" does not work; demo file doesn't work either; all give same error:

??? Error using ==> generation_TSP
Too many input arguments.

Error in ==> TSP_CE at 152
X = generation_TSP(Pt , option.N , X1);

Ye Ning

nino nona

allamul wafa

please send me the TSP solver where the input is coordinat

Carlos Alcántara

Updates

1.0.0.0

Missing cost_tsp.c file

-add missing files

-add a small procedure to detect two way path

v 1.5 : -add two mex files improving speed computation
-clean up the main CE_TSP function

v.1.4 : - Another trick leads to a O(n

v.1.3 : - Add random seed initialization procedure
- Add two Uniform generator - Smaller code - Hudge Speed improuvement (Over 1000% for large n)

- Add ULONG_NORM definition in generation_TSP.c
V.1.1: 10% faster, Smaller code, Less Memory consuming, Added view of state transition probabilities matrix during update

V.1.2: Small fixes, 4% faster, Smaller code, Less Memory consuming

- Add ULONG_NORM definition in generation_TSP.c
- 10% faster
- Smaller code
- Less Memory consuming
- Add view of the state transition probabilities matrix during update

MATLAB Release Compatibility
Created with R2007b
Compatible with any release
Platform Compatibility
Windows macOS Linux