Research Article
An Improved Ant Colony Optimization Based on an Adaptive Heuristic Factor for the Traveling Salesman Problem
| Input:, , , , , , , , , | | Output: | | (1) | Calculate Euclidean distance for the cities and get the number of cities as | | (2) | Initialize pheromone on all path with | | (3) | Apply k-means clustering for the cities (equations (7)–(10)) | | (4) | Separate non-classified cities (equation (11)) | | (5) | | | (6) | | | (7) | fordo | | (8) | ifthen | | (9) | | | (10) | | | (11) | else | | (12) | | | (13) | | | (14) | end | | (15) | fordo | | (16) | fordo | | (17) | ifthen | | (18) | ant i select next city using , (equations (12)–(13)); //special ant | | (19) | else | | (20) | ant i select next city by equation (2); //normal ant | | (21) | end | | (22) | Set tabu table for ant I | | (23) | end | | (24) | Calculate fitness of the corresponding solution obtained by ant i (equation (1)) | | (25) | end | | (26) | Select the best solutions for all ants (include normal ants and special ants) | | (27) | Update global best solution (optimal solution) | | (28) | Apply improved 2-opt algorithm to optimal solution (Section 3.3) | | (29) | Update pheromone on normal and special best solution separately (equations (14)–(16)) | | (30) | ifthen | | (31) | | | (32) | ifthen | | (33) | Re-initialize the pheromone on the global best solution with | | (34) | | | (35) | end | | (36) | else | | (37) | | | (38) | end | | (39) | end |
|
