Talk:Stochastic diffusion search

From Scholarpedia
Jump to: navigation, search

    This is a nice review of Stochastic Diffusion Search (SDS) one of swarm intelligence methods. The paper sets the context using ant algorithms as the closest 'relatives' of SDS. On general note it would probably benefit the reader to link it to a class of probabilistic algorithms considered in Computer Science community, most notably 'Go With the Winners' by Aldous and Vazirani.

    More specific comments:

    page 1, line 2 delete 'most'

    page 2, DIFFUSION section: rewrite to - 'Hypotheses are communicated between agents through inter-agent communication ...'

    CONVERGENCE - change occurencies of 'Criteria' to 'Criterion'

    page 3, 'a function only of' -> 'strongly dependent on'

    last sentence of Recruitment Strategies section -> change both 'an passive' to 'a passive'

    page 4, last sentence of (a): add ' in nonstationarly or dynamic problems' at the end of the sentence

    (b) 1st sentence of the 2nd paragraph : delete 'resides'

    Analysis, swap 'Markov' and 'Ehrenfest Urn models' in between first and secopnd bullet points.

    Personal tools

    Focal areas