An Improved Microcanonical Mean Field Annealing Algorithm

摘要

This paper proposed a new improved microcanonical mean field annealing algorithm (MMFA), and applied to the task scheduling. Firstly we put forward two new strategies, which are energy incentive strategy with sectioned and mixed energy compensation strategy. Secondly, we use the same energy function and new state generation method as MFA algorithm in the new MMFA algorithm, to ensure that the new state are transferred in the reduced direction of energy, thus to quicken the search speed and to improve the performance of the new MMFA. With the application of static tasks scheduling experiments on multi-processor, we had verified the excellence and progress of the MMFA algorithm.