Symbolic execution has been widely used in the field of software testing.The research shows that constraint solving is the most time-consuming part in symbolic execution,even though some optimization techniques are used.The speculative symbolic execution reduces the consuming time of constraint solved by making several continuous constraint solving merge into once.The success rate of every time guess is affected by the depth of conjecture and the direction of search,especially the direction of search.Therefore,how to guide the path search to conduct in the direction of success is very important to improve the efficiency of speculative symbolic execution.In this paper,ant colony algorithm was used to search the path.Firstly,according to the node condition,the weight of branch path was determined.Then,the weight of a branch was updated according to whether this branch is covered in every time guess.This paper chose the direction of search based on the weight of branch.The experimental results show that the proposed method can improve the efficiency of speculative symbol execution effectively.

The termination of loop programs plays an important role in program analysis.This paper focused on the termination of linear assignment loop programs which have no traditional linear ranking functions.Based on the precise computation of Aⁿx,this paper expanded the concept of traditionally defined raking functions,gave a definition of k-ranking functions and proved the existence of k-ranking functions.All the computations on the synthesis of k-ranking functions were done by Regularchains,a symbolic computation tool in Maple.Experimental results show that somelinearloop programs which have no traditional linear ranking functions indeed can be proven to be terminating by the proposed method.

Defect prediction based on code changes has the advantage of smaller code inspection cost,easy fault localization and rapid fixing.This paper firstly formalized this problem as a multi-objective optimization problem.One objective is to maximize the number of identified buggy changes,and the other objective is to minimize the cost of code inspection.There exist an obvious conflict between two objectives,so this paper proposed a novel method MULTI.This me-thod can generate a set of non-dominated prediction models.In the empirical studies,this paper chose six large-scale open source projects (with 227417code changes in total) and considerd ACC and POPT as evaluation indicators of perfor-mance.Final results show that the proposed method can perform significantly better than the state-of-the-art supervised methods (i.e.,EALR and Logistic) and unsupervised methods (i.e.,LT and AGE).