Utilization of simulated annealing to approximate the traveling salesman problem. Simulated annealing is an optimization technique that is a variant of the metropolis algorithm were temperature is treated as a variable, decreasing over time. This decrease in temperature increases the penalty for increasing the energy of a given state, making the algorithm less likely to move to a less probable state, and therefore more likely to range towards minima. When combined with MCMC's ability to explore the high probability regions of a given space, this allows the algorithm to shift from exploration to exploitation during operation.
jonmarty / sa-travelingsalesman Goto Github PK
View Code? Open in Web Editor NEWUtilization of simulated annealing to approximate the traveling salesman problem. Simulated annealing is an optimization technique that is a variant of the metropolis algorithm were temperature is treated as a variable, decreasing over time. This decrease in temperature increases the penalty for increasing the energy of a given state, making the algorithm less likely to move to a less probable state, and therefore more likely to range towards minima. When combined with MCMC's ability to explore the high probability regions of a given space, this allows the algorithm to shift from exploration to exploitation during operation.