Simulated annealing and Boltzmann machines: a stochastic approach to combinatorial optimization and neural computing
Journal of Computational Physics
Linear programming 1: introduction
Linear programming 1: introduction
Hard Instance Generation for SAT (Extended Abstract)
ISAAC '97 Proceedings of the 8th International Symposium on Algorithms and Computation
CSL '92 Selected Papers from the Workshop on Computer Science Logic
The complexity of theorem-proving procedures
STOC '71 Proceedings of the third annual ACM symposium on Theory of computing
Adaptive and scalable allocation of data-objects in the web
ICCSA'03 Proceedings of the 2003 international conference on Computational science and its applications: PartI
Golden ratio annealing for satisfiability problems using dynamically cooling schemes
ISMIS'08 Proceedings of the 17th international conference on Foundations of intelligent systems
Golden annealing method for job shop scheduling problem
MACMESE'08 Proceedings of the 10th WSEAS international conference on Mathematical and computational methods in science and engineering
Hi-index | 0.00 |
For Satisfiability (SAT) Problem there is not a deterministic algorithm able to solve it in a polynomial time. Simulated Annealing (SA) and similar algorithms like Threshold Accepting (TA) are able to find very good solutions of SAT instances only if their control parameters are correctly tuned. Classical TA's algorithms usually use the same Markov chain length for each temperature cycle but they spend a lot of time. In this paper a new hybrid algorithm is presented. This algorithm is in fact a TA algorithm which is hybridized with SA in a certain way. For this TA algorithm, the Markov chain length (L) is obtained in a dynamical way for each temperature. Besides, it is known that TA and SA obtain very good results whether their parameters are correctly tuned. Experimental tuning methods expend a lot of time before a TA algorithm can correctly be executed; in other hand, analytical tuning methods for TA were only completely developed for the geometrical cooling function. This paper also shows how TA can be tuned for three common cooling functions with an analytical model. Experimentation presented in the paper shows that the new TA algorithm is more efficient than the classical one.