Autor(es)

Venero, Sheila Katherine ; Schmerl, Bradley ; Montecchi, Leonardo ; dos Reis, Julio Cesar ; Fischer Rubira, Cecilia Mary

• Tipo de documento: Artigo Científico
• Fonte: ENTERPRISE, BUSINESS-PROCESS AND INFORMATION SYSTEMS MODELING, BPMDS 2020, EMMSAD 2020; v. 387, p. 16-pg., 2020-01-01.

Resumo

Knowledge-intensive Processes (KiPs) are processes characterized by high levels of unpredictability and dynamism. Their process structure may not be known before their execution. One way to cope with this uncertainty is to defer decisions regarding the process structure until run time. In this paper, we consider the definition of the process structure as a planning problem. Our approach uses automated planning techniques to generate plans that define process models according to the current context. The generated plan model relies on a metamodel called METAKIP that represents the basic elements of KiPs. Our solution explores Markov Decision Processes (MDP) to generate plan models. This technique allows uncertainty representation by defining state transition probabilities, which gives us more flexibility than traditional approaches. We construct an MDP model and solve it with the help of the PRISM model-checker. The solution is evaluated by means of a proof of concept in the medical domain which reveals the feasibility of our approach. (AU)

• Processo FAPESP: 19/02144-6 - Uma Metodologia para Apoiar o Reuso na Especificação de Modelos de Performabilidade
• Beneficiário: Leonardo Montecchi
• Modalidade de apoio: Auxílio à Pesquisa - Regular
• Processo FAPESP: 17/21773-9 - Desenvolvimento e Teste de Sistemas-de-Sistemas Resilientes: Uma Perspectiva de Arquitetura de Software
• Beneficiário: Cecília Mary Fischer Rubira
• Modalidade de apoio: Auxílio à Pesquisa - Regular