Memory embedded in Markov models specified in Statecharts: simulation versus analytical approaches

• Autor(a) principal: Vijaykumar,Nandamudi Lankalapalli
• Data de Publicação: 2012
• Outros Autores: Berkenbrock,Gian Ricardo, Carvalho,Solon Venâncio de, Andrade,Valéria Maria Barros de, Swamy,Gurrala Veereswara, Rao,Mokka Jagannadha
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Gestão & Produção
• Texto Completo: http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0104-530X2012000400001
• topic: Performance models; Statecharts specification; Memory representation; Analytical solutions; Continuous-time Markov chains; Simulation

Statecharts are a graphical representation to model reactive systems that respond to external or internal stimuli by changing the state of a given system. Statecharts can be seen as an extension of state-transition diagrams that allows modeling hierarchy, orthogonality, and interdependence. Due to their features to represent reactive systems, Statecharts have been adapted to represent and deal analytically with performance models (reactive systems whose performance is to be evaluated). An interesting feature present in Statecharts is to record the system's state, which cannot be represented in Markov models in a straightforward manner due to its "memory-less" property. The contributions of this paper are: show that Statecharts are a feasible alternative to specify a reactive system so that its performance can be evaluated by both analytical and simulation techniques; show that the inclusion of the memory representation in the Statecharts specification can indeed be made by both analytical and simulation techniques. The results of a case study of a manufacturing system show that the objectives are achieved.