tsp-simulated-annealing

Simulated annealing optimizer designed to solve Travelling Salesman Problem (TSP). Working of this optimizer consists of three stages:

• Stage 1: Greedy algorithm provides an initial solution,
• Stage 2: Simulated Annealing searches solutions space,
• Stage 3: 2-Opt algorithm improves the solution obtained in stage 2. (optional)

DEMO

• Input should be in the form of a list [ [x_1, y_1], [x_2, y_2], ... , [x_n, y_n] ]

from random import randint

towns = []

for town in range(60):
    towns.append([randint(0, 100), randint(0, 100)])

• How to initialize

from SimulatedAnnealing import SimulatedAnnealing

SA_optimizer = SimulatedAnnealing()
SA_optimizer.fit(towns, iterations=60000, cooling='polynomial', improve=True)

• Plotting results

SA_optimizer.plot()

• Visualization

SA_optimizer.show_graph(fitted=True)