Dynamic Scheduling Strategies for Firm Semi-Periodic Real-Time Tasks

This paper introduces and assesses novel strategies to schedule firm semi-periodic real-time tasks. Jobs are released periodically have the same relative deadline. Job execution times obey an arbitrary probability distribution can take either bounded or unbounded values. We investigate several optimization criteria, most prominent being Deadline Miss Ratio ( DMR ). All previous work uses some admission policies but never interrupt of admitted job before its On contrary, we introduce three new control parameters dynamically decide whether a at any given time. derive Markov model use stationary determine best value each parameter. Finally conduct extensive simulation campaign with 16 different distributions. The results nicely demonstrate how help improve system performance compared traditional approaches. In particular, show that (i) pre-execution rules, make significantly better decisions; (ii) specifically, key parameter is upper bound waiting time job; (iii) scheduling strategy decreases by up 0.35 over competitors.