A NOVEL EXPERIMENTAL ANALYSIS OF THE MINIMUM COST FLOW PROBLEM

摘要: In the GA approach parameters that influence its performance include population size, crossover rate and mutation rate. Genetic algorithms are suitable for traversing large search spaces since they can do this relatively fast because operator diverts method away from local optima, which will tend to become more common as space increases in size. GA’s based concept on natural genetic evolutionary mechanisms working populations of solutions contrast other techniques work a single solution. An important aspect is although not require any prior knowledge or limitations such smoothness, convexity unimodality function be optimized, exhibit very good most applications. The minimum cost flow problem formulated algorithm simulated annealing. This paper shows annealing much easier implement solving transportation problems compared with constructing mathematical programming formulations. Finally, new empirical study effect convergence SA demonstrated.