Jitter minimization with genetic algorithms

Transmission network induced jitter in periodic control variables is a known problem on field bus based distributed systems for embedded control applications. This jitter can be reduced or eliminated if adequate release instants are imposed to the periodic messages transmitted. In this paper, jitter reduction is achieved testing a variant genetic algorithm that determines an adequate initial phasing. This algorithm can be used progressively, i.e. it can work first on a subset of the messages, thus finding solutions for the higher priority ones, and later including the other messages until the whole set is considered. Experimental results obtained with two well-known and widely used benchmarks, the PSA, coming from automotive industries, and the SAE from Automatically Guided Vehicles, show complete elimination or a significant decrease of jitter when compared to non-optimized systems.