A Parallel Simulated Annealing Processor for Real-Time Systems

Simulated Annealing is a well-known optimization algorithm for arbitrary combinatorial problems. The drawback of this algorithm is its poor performance. The goal of this research project is to develop a special purpose processor for simulated annealing. This processor should be able to handle time constraints so that it can be used in real-time systems. The main features of this processor are: Massively parallel computation to decrease computation time, continuous status information for use in a real time environment, the possibility to interrupt computation and continue with the best solution achieved up to that point, and an easy programming environment. This processor will be used for online task-mapping, HW/SW codesign and other NP-hard problems in computer science.