This chapter deals with Reconfigurable Uniprocessor embedded Real-Time Systems to be classically implemented by different OS tasks that we suppose independent, asynchronous, and periodic in order to meet functional and temporal properties described in user requirements. The authors define a schedulability algorithm for preemptable, asynchronous, and periodic reconfigurable task systems with arbitrary relative deadlines, scheduled on a uniprocessor by an optimal scheduling algorithm based on the EDF principles and on the dynamic reconfiguration. Two forms of automatic reconfigurations are assumed to be applied at run-time: Addition-Remove of tasks and just modifications of their temporal parameters: WCET and/or Periods. Nevertheless, when such a scenario is applied to save the system at the occurrence of hardware-software faults, or to improve its performance, some real-time properties can be violated. The authors define a new semantic of the reconfiguration where a crucial criterion to consider is the automatic improvement of the system’s feasibility at run-time by using an Intelligent Agent that automatically checks the system’s feasibility after any reconfiguration scenario to verify if all tasks meet the required deadlines. Indeed, if a reconfiguration scenario is applied at run-time, then the Intelligent Agent dynamically provides otherwise precious technical solutions for users to remove some tasks according to predefined heuristic (based on soft or hard task), or by modifying the Worst Case Execution Times (WCETs), periods, and/or deadlines of tasks that violate corresponding constraints by new ones, in order to meet deadlines and to minimize their response time. To handle all possible reconfiguration solutions, they propose an agent-based architecture that applies automatic reconfigurations in order to re-obtain the system’s feasibility and to satisfy user requirements. Therefore, the authors developed the tool RT-Reconfiguration to support these contributions that they apply to a Blackberry Bold 9700 and to a Volvo system as running example systems and we apply the Real-Time Simulator Cheddar to check the whole system behavior and to evaluate the performance of the algorithm (detailed descriptions are available at the Website: http://beru.univ-brest.fr/~singhoff/cheddar). The authors present simulations of this architecture where they evaluate the agent that they implemented. In addition, the authors present and discuss the results of experiments that compare the accuracy and the performance of their algorithm with others.
