xavier-mayiming / tabu-search

Tabu search for traveling salesman problem (TSP)

Tabu Search

Reference: Glover, Fred. Tabu Search-- Part I.[J]. ORSA Journal on Computing, 1989.

The tabu search for the traveling salesman problem

Variables	Meaning
x	The x coordinates of cities (list)
y	The y coordinates of cities (list)
iter	The maximum number of iterations
city_num	The number of cities
na	The number of actions
dis	The distance matrix, dis[i][j] denotes the distances between city i and city j
TC	The tabu counter (list)
TL	The tabu length
gbest	The length of the global best path
gbest_path	The global best path (list)
iter_best	The length of the global best path of each iteration (list)
con_iter	The last iteration number when "gbest" is updated

Example

if __name__ == '__main__':
    min_coord = 0
    max_coord = 10
    city_num = 30
    iter = 50
    x = [random.uniform(min_coord, max_coord) for _ in range(city_num)]
    y = [random.uniform(min_coord, max_coord) for _ in range(city_num)]
    print(main(x, y, iter))

Output:

The TS converges at its 20-th iteration, and the global best value is 46.273638987858966.

{
    'best length': 46.273638987858966, 
    'best path': [16, 6, 15, 20, 24, 25, 5, 21, 12, 22, 18, 28, 14, 4, 0, 26, 2, 27, 7, 10, 9, 17, 11, 19, 23, 3, 29, 13, 1, 8], 
    'convergence iteration': 20
}