Analiza parametrów czasowych w systemach pomiarowo-steruj?cych z wykorzystaniem teorii masowej obs?ugi i metod szeregowania zadań

摘要

In currently designed measurement and control systems (MCS) with realtime constrains, the task scheduling theory is used to calculate the tasks response time [3, 4]. Such approach is suitable for hard real-time MCS and sometimes can lead to oversizing the resources of the designed MCS. For design of soft real-time MCS a better solution can be to apply the queuing theory. The queuing theory is based on mean values of a task execution time and times between the successive tasks. Such approach allows achieving better allocation of MSC resources for task execution but sometimes tasks deadlines can be exceeded. In the paper a general introduction to the task scheduling theory and the queuing theory is presented. The basic time parameters of a task in both approaches are the task execution time and the task period occurrence. In Paragraphs 4 and 5 the method for calculation of the utilization-based coefficient for the task scheduling theory (Formula 2) [3] and the rate of occupancy of service in the queuing theory are outlined (5) [7]. Formulas 3 and 4 describe the task response time for a system with or without pre-emptive system determined with use of the task scheduling theory [3]. In Paragraph 5 calculation of the mean task residence time in the non-pre-emptive (7) and pre-emptive system (10), as well as the mean task residence time in the queue with use of the queuing theory are presented (formulas 6 and 8) [5, 7], The results of carried out calculations and simulations contained in Paragraph 6 show usability of the queuing theory for analysis of MCS.%W artykule zaprezentowano mo?liwo?ci wykorzystania teorii masowej obs?ugi w analizie systemów pomiarowo-steruj?cych (SPS). Wyniki zosta?y zestawione z metod? szeregowania zadań oraz wynikami symulacyjnymi. Analizie poddano parametry zwi?zane z warunkiem realizowal-no?ci zadań, ?rednim i maksymalnym czasem wykonania zadań oraz ?rednim czasem oczekiwania zadania w kolejce. Dodatkowo wyniki zosta?y potwierdzone poprzez symulacj? wykonan? w programie NS2 (Network Symulator).