Research on the Initial Value of the Simulated Annealing

Abstract:

Simulated Annealing Algorithm is one of the top ten classical optimization algorithm, and it has been successfully applied to various fields. Simulated annealing is a optimization algorithm which can find the global optimal solution, compares to neural network algorithm, it is so easily to implement that has higher probability to be adopted, but it has own shortcomings like other optimization algorithms, its result largely depends on initial value, The initial value of the traditional simulated annealing algorithm began with a random number, its convergence speed is often slow very much and the effect is bad. In this paper, a new simulated annealing algorithm that based on genetic algorithm and hill-climbing method was brought up, because of hill-climbing algorithm was easy to fall into local optimum, and simulated annealing can just solve the problem, it not only escaped from local optimum, but also got good convergence speed and results.

• Periodical: Advanced Materials Research (Volumes 774-776)
• Pages: 1770-1773
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.774-776.1770
• Online since: September 2013
• Authors: Xiao Wen Liang, Wei Gong, Wen Long Fu, Jing Qi
• Keywords: Genetic Algorithm (GA), Hill-Climbing Method, Initial Value, Simulated Annealing Algorithm

References

[1] Haibo Liao, Zhongying Wan, and Mingwen Wang: Projection Pursuit model of simulated annealing in the application of text classification, pp.628-635.

[2] Toby Segaran: Programming Collective Intelligence, (2009).

[3] Haodong Zhu, and Yong Zhong: Using optimization simulated annealing algorithm for text feature selection, edited by Computer engineering and applications, Vol. 46(4) (2010), pp.613-620.

[4] Deck G, Scheuer T: Threshold accepting: A general purpose optimization algorithm appearing uperior to simulated annealing, submitted to Journal of Computational Physics, Vol. 90(1) (1990), pp.161-175.

DOI: 10.1016/0021-9991(90)90201-b