A novel Tree-coding Genetic Algorithms (TGA) is presented for QoS-aware service composition. Since Tree-coding schema can carry the information of static model of service workflow, this feature qualifies TGA to make the chromosomes to be encoded and decoded automatically, and keep the medial result for fitness computing. The Tree-coding can also support the services composition re-planning at runtime effectively. The experiment results show that TGA run faster than the one-dimensional coding GA when the optimal result is same, furthermore the algorithm with Tree-coding is effective for re-planning.
We introduce an improved version of the symbolic transition graph with assignment (STGA) of Lin. The distinction of our model is that the assignment of a transition is performed after rather than before the action. Consequently, it has two advantages over the original one: on one hand, most regular value-passing processes can be represented more intuitively and compactly as such graphs; on the other hand, the natural definitions of symbolic double transitions can be given. The rules which generate the improved STGAs from regular value-passing processes are presented. The various versions (late/early, ground/symbolic) of strong operational semantics and strong bisimulation are given to such graphs, respectively. Our strong bisimulation algorithms are based on the late strong bisimulation algorithm of Lin, however, ours are more concise and practical. Finally, the improved STGAs are generalized to both symbolic observation graphs with assignments and symbolic congruence graphs with assignments, and therefore weak bisimulation equivalence and observation congruence can be checked, respectively.
Abstract. For ensuring the displaying effects of Web applications, it is important to perform compatibility testing for browsers under the different configurations and it is a hard task to test all. So this paper focused on the improvements for browser compatibility testing. Firstly, we introduced the related work. Then we analysed the functional speciality of the current popular browsers, and provided the preconditions to simplify problems. Next we gave an instance and brought forward the single factor covering method and the pair-wise covering design to gain testing suits. Finally, we introduced the assistant tools we have developed and the future work.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.