EFSM executable test sequence generation method based on Monte Carlo tree search

Abstract

The invention discloses an EFSM executable test sequence generation method based on Monte Carlo tree search. The methd comprises the steps of converting an executable test generation problem of the EFSM model into a Markov decision process of a path in a TEA tree. According to the method, a target sequence is searched by heuristically guiding the extension direction of a TEA tree by means of the powerful massive state space search capability of an MCTS algorithm, so that the test generation efficiency is greatly improved, and finally state explosion is prevented; therefore, the method disclosed in the can effectively reduce the expansion nodes of the executable analysis tree and accelerate the generation of the executable test sequence.

Description

technical field [0001] The invention belongs to the technical field of protocol consistency test sequence generation, and in particular relates to an EFSM executable test sequence generation method based on Monte Carlo tree search. Background technique [0002] The increasing complexity of communication networks makes it increasingly difficult to develop and test communication protocols. Subtle errors in the communication protocol software may also cause anomalies in the entire communication network, resulting in immeasurable losses. In order to improve the quality of communication protocol software, it is an effective way to carry out detailed network protocol conformance testing. As an important branch of protocol testing, network protocol conformance testing mainly aims to detect the degree of conformance between protocol implementation and protocol specification; Model-Based Testing (MBT) based on formal model helps to improve the testing performance. The quality and d...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the executable problem of test sequence, researchers have carried out a lot of exploration work, among which the search technology based on genetic algorithm has become a main camp, they transform the generation of EFSM test sequence into an optimization problem, use genetic algorithm to search and solve, The most representative one is the method proposed by Kalaji "Kalaji, A.S., R.M.Hierons, and S.Swift, An integrated search-based approach for automatic testing from extended finite state machine (EFSM) models. Information and Software Technology, 2011.53(12) :p.1297-1318.", this method first uses an algorithm to generate candidate test sequences that meet the coverage requirements, but these test sequences are not necessarily executable, and then uses GA to generate test data that can trigger the candidate test sequences to be executable ; However, there are difficulties in initial population selection and population length setting in this type of method, and the feasibility of the generated test sequence still needs additional steps to further verify and confirm

Some researchers also propose to generate test sequences based on constraint solving and symbolic execution techniques, but these methods have certain constraints on the model under test and are not universal

[0005] Compared with the above methods, Transition Executable Analysis (TEA) is more universal. It mainly generates the target test sequence through the extended search of Transition Executable Analysis Tree. However, for the actual protocol, the huge reachable state space of EFSM corresponds to The massive potential expansion search space of the executable parse tree of , eventually leads to the state explosion problem

Images