Abstract: Test case generation based on Web application front-end model is an important process of Web application testing,but most existing models for Web applications only focus on Web pages and their events,ignoring event triggering conditions and subsequent actions.Therefore,in order to describe the dynamic behavior of modern Web applications more accurately,this paper defines a new Web application front-end behavior model(FBM).Because there may be internal variables in the triggering conditions of transition in the model,that is,there are dependencies between transitions,which will make the generated test cases cannot be executed according to the input sequence,thus affecting the test results.Therefore,an optimized grouping genetic algorithm is proposed to automatically generate the feasible transition path(FTP).Considering the characteristics of FTP generation problem,the algorithm makes a reasonable design of chromosome initialization and fitness function,and adds a repair operator to adjust the individual length to generate FTP which satisfies the migration coverage.This paper also introduces an adaptive genetic operator and simulated annealing receiving mechanism to reduce the number of iterations,thus improving the solution speed.Experimental results show that the algorithm can effectively guarantee the feasibility and coverage of transition path on the basis of higher solution efficiency.

Key words: Web application testing, Front-end behavior model, Feasible test case generation, Grouping genetic algorithm