Abstract:A fast simulated annealing algorithm has been formed by using both the generalized Boltzmann-Gibbs statistics and the turbulent model derived from temperatore-dependent pseudo Cauchy distribution.The formulae for determining the acceptance probability and temperature lowering way under such conditions are given respectively.The new algorithm achieves big scope search at high temperature and minor scope search around the model at low temperature.The smooth"tail"of pseudo Cauchy distribution favours avoiding the restriction of local extremums,so that the simulated annealing alogorithm converges very fast.