where , τij(t) is the residual pheromone on the path between node i and node j at moment t. m is the number of ants.Cmin is the minimum of product of distance and traffic density between the two junctions. In the process of planning path,the kth(k=1,2,...,m) ant chooses next junction based on probability
S = [τij(t)]. [ij] …show more content…
Using Arrangement of the Ant Colony Algorithm to solve Chinese TSP problem setting m=25, α=1, β = 5, ρ = 0.25, the number of ant that can release pheromone is 7, and the maximum iterations is 100, the simulation results show that its optimal solution is 15,381km, and the iterations is only 58 times[10]. Which is less than genetic algorithm,alignment algorithm and other algorithms.But it still falls into local optimum easily and has low probability to get the optimal …show more content…
The impact on ant selection for next iteration is greater, it is easier to get the same solution, so the solution is stable, but it is not the optimal solution.
Inspired by this,we improve the ant colony algorithm with introducing dynamically adjust ant number. We neither consider too large ant number nor too small. increaseThis increases the global search capability to find the optimal result.We can update ant number in two ways. one is updated fixedly and the other way is updated randomly.We start with fewer ants at beginning, and the ant number is updated with the iteration process,where each time the updation number is same.Whereas,in update-randomly way,update number is random.The flow chart of this improved version of algorithm is as