A genetic algorithm based configuration optimization method for AFDX

Abstract

The configuration table of AFDX(Avionics Full Duplex Switched Ethernet) network is used for organizing the message transmission and setting the path for transmission. Its design must satisfy the timeliness of data transmission. However, existing design of configuration table causes too much delay on AFDX network. To fix this problem, this paper firstly provides a two-layer optimized method for the network max frame length of AFDX network configuration table. The method ensures the timeliness of high priority VL (virtual link), reserves more bandwidth for low priority VL and optimizes the end-to-end delay of low priority VL. Secondly, the method is solved with Genetic Algorithm to obtain the optimized max frame length. Finally, a test case is designed which shows the effectivity and practicality of the algorithm. This study not only ensures the timeliness and reliability of AFDX data transmission, but also optimizes the allocation of resources and improves the efficiency of configuration loading.