the weight vector of the graph is multiplied in parallel by the pre-processed matrix to obtain the best Hamiltonian path. Figure 3 shows three columns, with each column representing a city.