A parallelizable operation scheme of the Boltzmann machine optimizer based on group updates
Aristidis C. Likas, George J. Papageorgiou, Andreas Stafylopatis · 1995
An operation scheme for the Boltzmann Machine Optimizer is proposed thatis suitable for parallelization and is based on the notion of group updates. It hasthe advantage of suggesting transitions between states that differ in more than oneunit and exhibits greater flexibility in accepting such transitions compared to thepure sequential case. The performance of the method is evaluated on the WeightedMaximum