Hybrid Metaheuristic Approach for Scheduling of Aperiodic OS Tasks - Complex System Modelling and Control Through Intelligent Soft Computations - page 834

Experimental Results: Discussion and Evaluation

In this subsection, in order to quantify the bene ts of the proposed approach over

the other works from the state of the art and in order to check also the suggested
con gurations of tasks allowing the system ' s feasibility and the response time

minimization, we simulate the agent s behavior on several test sets in order to rate

the performance of the optimal scheduling algorithm in our proposed approach. The
most important observation was obtained by the comparison of our proposed
approach against the others from the literature about the current values. We tested
the feasibility of the Volvo task system by another algorithms, so that we can
compare the results directly. We carried out several test runs and examined them
under different aspects. The total utilization of the static schedule is 75 %, the
classic one is 145.4 % and the other proposed by our method is 63.1 %.
Therefore, we can con

'

rm that this method is nowadays very advantageous
given the fast response time and the performance of the RT-Recon guration tool.

By applying the six solutions of this tool RT-Recon guration, we conclude that

our approach can allow more reactive and also more ef cient feasible systems. This

advantage can be important in many cases where critical control tasks should be
intensively executed in small periods of time. This work also, concentrates on the
context of systems containing a set of tasks which is not feasible; the recon gu-

ration was applied in order not only to obtain the system s feasibility but also to get

the performance of the system by reducing the response time of the processes to be
tolerated in interactive environment and by avoiding unnecessarily frequent context
switch leading to more overheads resulting in less throughput. This advantage was
increased and proved clearly by the volvo case study. This advantage was illus-
trated by examples even for the case of aperiodic tasks also.
Moreover, with the revolution of semiconductors technology and the develop-
ment of ef

'

cient recon guration tools, the use of our method and the RT-Recon-

figuration tool will become increasingly important, and very advantageous for rapid
and ef gurable OS tasks,

especially when the user has no other choice than to choose the previous proposed
solutions and to decide the proper values of each recon

cient response time of the aperiodic/sporadic recon

gured task ' s parameters in

order to obtain the system ' s feasibility and to minimize the response time of the

studied systems.

4.1.2 Recon guration of Aperiodic Tasks

The scheduling problem for aperiodic tasks in real-time embedded systems is very
different from that for periodic tasks. Scheduling algorithms for aperiodic tasks
must be able to guarantee the deadlines for hard aperiodic tasks (also called a
sporadic tasks) and provide good average response times for soft aperiodic tasks.
Our aperiodic scheduling approach must also accomplish these goals without
compromising the hard deadlines of the periodic tasks in the case when we have a
mixture of periodic, sporadic and aperiodic tasks. We assume also in our original