Hybrid Metaheuristic Approach for Scheduling of Aperiodic OS Tasks - Complex System Modelling and Control Through Intelligent Soft Computations - page 815

1 Introduction

The use of computers for control and monitoring of industrial processes has
expanded greatly in recent years, and will probably expand even more dramatically
in the near future Layland and Liu ( 1973 ). Indeed, Real-time systems are used to
control physical processes that range in complexity from automobile ignition sys-
tems to controllers for flight systems and nuclear power plants. In these systems, the

correctness of system functions depends upon not only the results of computation
but also on the times at which results are produced. Also, a process control com-
puter performs one or more control and monitoring functions. The IEEE Real-Time
Systems Symposium has in the last two decades been the main forum for publishing
the key results in aperiodic real-time scheduling theory that are reviewed in this
topic chapter work. A number of other initiatives were funded also including a
series of in

fl

uential studies commissioned by the European Space Agency. For
example, the works appear in Abdelzaher et al. ( 2004a , b ); Abeni and Buttazzo
( 1998 , 2001 , 2004 ); Aggarwal and Chraibi ( 1995 ). Various scheduling policies for
real-time multiprocessors systems have been constantly evolving. Scheduling of
real-time applications/tasks on multiprocessor systems often has to be combined
with task to processor mapping. With the current design trends moving towards
multicores and multiprocessor systems for high performance and embedded system
applications, the need to develop design techniques to maximize utilization of
processor time, and at the same time minimize power consumption, have gained
importance. General design techniques to achieve the above goals have been fueled
by the development of both hardware and software solutions. Hardware solutions to
minimize power and energy consumption include: dynamic voltage and frequency
scaling (DVFS) processors, dynamic power management modules Advanced
Con

fl

guration and Power Interface (ACPI), thermal management modules, intelli-
gent energy management (IEM), and heterogeneous multicore or multiprocessor
systems. Software solutions for maximizing processor utilization and energy min-
imizations include: parallelization of instructions, threads, and tasks; effective
thread/task scheduling; and mapping algorithms for multicore and multiprocessor
systems Naveen and Venkatesan ( 2013 ). In Layland and Liu ( 1973 ), the authors
used the example of the antenna pointing to track a spacecraft in its orbit. In this use
case, each function to be performed has associated with it a set of one or more tasks
where some of these tasks are executed in response to events in the equipment
controlled by or monitored by the computer. The remainder are executed in
response to events in other tasks and none of the tasks may be executed before the
event which requests it occurs and each of them must be completed before some

fixed time has elapsed following the request for it. This kind of tasks is called a real-
time tasks. A real-time task is generally placed into one of four categories based
upon its arrival pattern and its deadline. If meeting a given task

' is deadline is critical

to the system ' is operation, then the task ' is deadline is considered to be hard. If it is

desirable to meet a task s deadline but occasionally missing the deadline can be

tolerated, then the deadline is considered to be soft. Tasks with regular arrival times

'