Scilab Home page | Wiki | Bug tracker | Forge | Mailing list archives | ATOMS | File exchange
Please login or create an account
Change language to: English - Português - 日本語

Please note that the recommended version of Scilab is 6.0.1. This page might be outdated.
See the recommended documentation of this function

Aide Scilab >> Simulated Annealing > compute_initial_temp


A SA function which allows to compute the initial temperature of the simulated annealing

Calling Sequence

T_init = compute_initial_temp(x0,f,proba_init, ItMX [, param] )



the starting point


the objective function which will be send to the simulated annealing for optimization


the initial probability of accepting a bad solution (usually around 0.7)


the number of iterations of random walk (usually around 100)


optional, a data structure managed with the parameters module.

The optim_sa function is sensitive to the following fields.


a function which computes a neighbor of a given point. The default neighbourhood function is neigh_func_default.


the type of acceptation function. If the type is equal to "sa", then the initial temperature is computed from T_init = - f_sum ./ log(proba_init). If the type is equal to "vfsa", it is computed from T_init = abs(f_sum / log(1/proba_init - 1)).


The initial temperature corresponding to the given probability of accepting a bad solution


This function computes an initial temperature given an initial probability of accepting a bad solution. This computation is based on some iterations of random walk.


x0 = [2 2];
Proba_start = 0.7;
It_Pre = 100;
x_test = neigh_func_default(x0);

saparams = init_param();
saparams = add_param(comp_t_params,'neigh_func', neigh_func_default);

T0 = compute_initial_temp(x0, rastrigin, Proba_start, It_Pre, saparams);

See Also

  • optim_sa — A Simulated Annealing optimization method
  • neigh_func_default — A SA function which computes a neighbor of a given point
  • temp_law_default — A SA function which computed the temperature of the next temperature stage


Yann Collette - 2008 -

Michael Baudin - Digiteo - 2010

Scilab Enterprises
Copyright (c) 2011-2017 (Scilab Enterprises)
Copyright (c) 1989-2012 (INRIA)
Copyright (c) 1989-2007 (ENPC)
with contributors
Last updated:
Wed Oct 05 12:11:00 CEST 2011