On Test Sequcence Generation of Time-Triggered Ethernet Using Genetic Algorithm

The model testing method in software testing theory can effectively reduce the uncertainty in testing, increase the effectiveness of test sequence generation, and lay the foundation of automated testing process. The Extended Finite State Machine (EFSM), being a tool that can precisely describe the dynamic process of a software or protocol et.al, has been widely studied in the field of software testing. Different from the traditional Finite State Machine (FSM), EFSM additionally introduced data flow process on the basis of the original control flow, which however limits ability to generate Feasible Transition Paths (FTP). As a kind of heuristic algorithm, Genetic Algorithm (GA) is able to take control flow and data flow simultaneously into account to ensure the efficiency of test sequence and test case generation process. Being first to apply GA to TTE, this paper selects the compression master (CM) protocol state machine from the Time-Triggered Ethernet (TTE) clock synchronization protocol, and uses a genetic algorithm to generate executable test paths for the state machine, tending to find a new way to study the test case generation of TTE. In the process, a single-objective optimization method will be used, and at the end of the essay, a comparation of a GA-generated test path between a randomly generated test path will be proceed to conclude that it is an efficient way.