Application of the “Winner Takes All” Principle in Wang’s Recurrent Neural Network for the Assignment Problem

A Recurrent Neural Network to Traveling Salesman Problem

“…The method considered in this work uses one technique based on the "Winner Takes All" principle, speeding up the convergence of the Wang's Neural Network, besides correcting eventual problems that can appear due the multiple optimal solutions or very closer optimal solutions (Siqueira et al, 2005).…”

“…In all the tested matrices, η = 1 had been considered, and parameters τ and λ had been calculated in many ways, described as follows (Siqueira et al, 2005).…”

A New Heuristic Procedure To Solve The Traveling Salesman Problem

“…The heuristic technique proposed in this work uses the WTA principle, accelerating the WRNN's convergence and correcting eventual problems that may appear due to multiple optimal solutions or optimal solutions that are very close to each other [17].…”

“…The parameters chosen for the WRNN are those that determine the best results for the AP [17]. Parameter η is considered equal to 1 in all cases tested within this work.…”

“…The Wang Recurrent Neural Network (WRNN) with the "Winner Takes All" (WTA) principle [17] can be applied to solve the TSP, solving this problem as if it were an Assignment Problem (AP) by means of the WRNN, and, furthermore, using the WTA principle on the solutions found with the WRNN, with the constraint that the solutions found must form a feasible route for the TSP. The parameters used for the WRNN are those that show the best solutions for the AP.…”