Plenary lecture II: annealing hybrid algorithms strategies for NP class problems
Juan Frausto–Solís · International Conference on Mathematical and Computational Methods in Science and Engineering · 2008
One of the main objectives of Computer Science is to develop algorithms for helping human beings to solve their difficult and important problems with speed and quality. Among the more difficult computational problems are those belonging to the NP Hard class; examples of these problems are: Satisfiability Problem (SAT), Scheduling (including the allocation of tasks in operating systems and robotics), planning and most problems related with bioinformatics such as phylogenetic trees construction, Folding problems and many others. There are many stochastic approaches proposed for these problems but none of them is always the best solution. A very good approach is to use Simulated Annealing hybridised (i.e. mixed) with other approaches. Among these approaches we can find the very formal ones as Mechanical Statistical, Markov Models, semi-formal as Support Vector Machines and Neural Networks, and very informal as Golden Ratio. In this presentation the main hybridization approaches, applications and challenges for future research are presented.