Microcanonical Annealing Algorithm with Energy Award Strategy
Xin Zhan-hong · Journal of Chinese Computer Systems · 2008
In order to improve the solution quality of traveling salesman problem tackling with microcanonical annealing algorithm,an energy award strategy applied on demon was proposed.It enhanced the chances of escaping from local optimum by augmenting the energy taken by demon when rejecting a possible state.Two kinds of implementations were presented,one is energy award without upper bound and the other is that with upper bound.The simulations showed that the former improved the solution qualities successfully while the latter failed on the benchmark.To quicken the convergence of the new strategy without upper bound,a tolerance mechanism was designed to attain a compromise between solution quality and convergence speed and its effect was demonstrated by the simulations.