Browsing by Subject "distributed graph transformations"
Now showing 1 - 2 of 2
- Results Per Page
- Sort Options
Item type:Article, Access status: Open Access , Collective agent strategies in the GRADIS environment(Wydawnictwa AGH, 2009) Sędziwy, AdamThe common obstacle in using a graph representation of the problems, used successfully in many areas (e.g. pattern recognition) is time complexity of related computations. In some cases only using the parallel computations applied to the graph grammars with quadratic parsing time make the graph approach applicable in practical use. The paper focuses on the multiagent methods of obtaining the graph partitions in the GRADIS framework, that are optimal for further effective parallel computations (decomposing of a graph representation). Various criteria of optimum are discussed. Also the results of performed tests are presented.Item type:Article, Access status: Open Access , Środowisko agentowe w syntaktycznym rozpoznawaniu obrazów(Wydawnictwa AGH, 2010) Sędziwy, AdamA computational complexity of parsing is the major limitation of syntactic pattern recognition methods applicability. That complexity may be decreased only by using grammars having weaker expressive power. ETPL(k) graph grammars, having quadratic complexity, are a compromise between complexity and descriptive power. For problems having large sizes however it's necessary to parallelize computations by distributing them over a multiagent system. Effectiveness of a system may be additionally boosted by even knowledge distribution enabling a balanced workload of agents. The knowledge consists of given subgraphs and grammar productions.
