On efficiency of one local algorithm for solving the traveling salesman problem

The well-known local steepest descent algorithms for solution of the traveling salesman problem are compared with what Sarvanov and Doroshko propose [1]. With random choice of the original Hamiltonian loop the latter algorithm finds in the vicinity of that loop a shorter Hamiltonian loop than do the conventional local steepest descent algorithms for nearly all complete symmetrical digraphs with some discrete weights of arcs.