Hybrid Metaheuristic Approach for Scheduling of Aperiodic OS Tasks - Complex System Modelling and Control Through Intelligent Soft Computations - page 843

Fig. 4 The simulation with sporadic server and background server

guration scenario is applied at t1 and t2 time units with
the arrival of 2 new aperiodic tasks e 1 at t1 = 7 and e 2 at t2 = 11 time units.
Therefore the system is feasible by applying the polling server to schedule the
system but the response time is equal to 17 and 33 for both e 1 and e 2 respectively.
Now by applying our new hybrid approach, the response time of the second arrival
aperiodic task is decreased from 33 to 25 time units as we observe in Fig. 2 .By
applying the new hybrid approach with Deferrable Server and Background server,
the response time of the

We suppose that a recon

first arrival aperiodic task is decreased from 17 to 10 time
units and the response time of the second arrival aperiodic task is decreased from 33
to 25 time units as we observe in Fig. 3 . Finally, by applying the new hybrid
approach with Sporadic Server and Background server, the response time of the

first arrival aperiodic task is decreased from 17 to 16 time units and the response
time of the second arrival aperiodic task is decreased from 33 to 24 time units as we
observe in Fig. 4 .

4.2.3 Formalization

By considering real-time operating system (OS) tasks scheduling, let n = n 1 + n 2 be
the number of a mixed workload of periodic and aperiodic tasks in Current ʓ (t).

The recon guration of the system Current ʓ (t) means the modi cation of its

implementation that will be as follows at t time units:

Current C t ðÞ ¼ n new [ n old

where n old is a subset of n 1 old periodic tasks which are periodic and not affected by

the recon guration scenario (e.g. they implement the system before the time t), and

n new is a subset of n 2 new aperiodic tasks in the system. We assume that an updated