Análise sob restrições baseada em um raciocínio energético para o problema de escalonamento tempo real de sistemas de gerenciamento de workflow

Abstract

This work propose a method of analysis under constraints applied to the problem of real time scheduling in workflow management systems. The goal is to filter the start of activities in order to eliminate inconsistent dates with scheduling solutions and to identify inadmissible schedules due to the inability of energy resources involved. The adopted model is a p-time Petri net with hybrid resource allocation mechanism. The proposed method uses originally mechanism of time constraint propagation based on cases duration and dierent types of routings for workflow processes. The diflerent types of resources (discrete and continuous) are incorporated into the model and an approach based on energetic reasoning is applied. On the set of modified constraints is then applied a specialized inference mechanism, called token player, seeking to obtain in real time an admissible scenario corresponding to a specific sequence of activities which respects all the time constraints.