60 likes | 292 Views
Solving Traveling salesman Problem with Hopfield Net. Aiming Lu. Traveling Salesman Problem.
E N D
Solving Traveling salesman Problem with Hopfield Net Aiming Lu
Traveling Salesman Problem • Given a finite number of cities along with the cost of travel between each pair of them, find the cheapest way of visiting all the cities and returning to your starting point. Every city can only be visited only once.
Apply Hopfield net to TSP • Use an n by n matrix to represent a tour. Vxi– x-th city as the i-th stop. • The problem is to construct an energy function that can be used to find a stable state of the network that expresses the cheapest valid city tours.
Energy Function E1: Row inhibition, favor only 1 city in a row E2:Col inhibition, favor only 1 city in a col E3: Global inhibition, favor the state that all cities are present E4:Distance inhibition, favor minimum distance of the tour A,B,C,D,N: are constants
Learning Procedure • Initiation • Cities distance • uxi(0) = u00 + ((rand-1)/10.*u0) • Repeat iteration until stop criterion is satisfied: uxi(t+1) = uxi(t) + t(duxi/dt) Vxi=(1+tanh(uxi/u0)