Approximate graph matching for software engineering

Approximate Graph Matching Many practical problems can be modeled as approximate graph matching (AGM) problems in which the goal is to find a ”good” matching between two objects represented as graphs. We first select, out of the possible formulations, the Error Tolerant Graph Matching (ETGM) framework, which is able to model most AGM formulations. Given that AGM problems are generally NP-hard, we based our resolution approach on meta-heuristics, given the demonstrated efficiency of this family of techniques on (NP-)hard problems. Our approach avoids as much as possible assumptions about graphs to be matched and tries to make the best out of basic graph features such as node connectivity and edge types. Consequently, the proposal is a local search technique using new node similarity measures derived from simple structural information. The proposed technique was devised as follows. First, we observed and empirically validated that initializing a local search with a very small subset of ”correct” node matches is enough to get excellent results. Thus, instead of directly trying to correctly match all nodes and edges from two graphs, one could focus on correctly matching a few nodes. Second, in order to retrieve such node matches, we resorted to the concept of local node similarity which consists in analyzing nodes’ neighborhoods to assess for each possible node match the likelihood of its inclusion in a good matching. We investigated many ways of computing similarity values between pairs of nodes and proposed additional techniques to attach a level of confidence to computed similarity value. Our work results in a similarity enhanced tabu algorithm (Sim-T) which is demonstrated to be more accurate and efficient than known state-of-the-art algorithms. Approximate Diagram Matching in software engineering Given the size and complexity of OO systems, retrieving and understanding the history of the design evolution is a difficult task which requires appropriate techniques. Building on the work done for generic AGM problems, we propose MADMatch, a Many-to-many Approximate Diagram Matching algorithm based on an ETGM formulation. In our approach, design representations are modeled as attributed directed multi-graphs. Transformations such as modifying, renaming, or merging entities in a software diagram are explicitly taken into account through edit operations to which specific costs can be assigned. MADMatch fully integrates the textual information available on diagrams and proposes several concepts enabling accurate and fast computation of matchings. We notably integrate to our proposal the use of termal footprints which capture the lexical context of any given entity and is exploited in order to reduce the search space of our tabu search. Through several case studies involving different types of diagrams (such as class diagrams, sequence diagrams and labeled transition systems), we show that our algorithm is generic and advances the state of art with respect to scalability and accuracy. Design Evolution Metrics for Defect Prediction Testing is the most widely adopted practice to guarantee reasonable software quality. However, this activity is often a compromise between the available resources and sought software quality. Following the retrieval of class diagrams’ evolution by our graph matching approach, we proposed and investigated the usefulness of elementary design evolution metrics in the identification of defective classes. The metrics include the numbers of added, deleted, and modified attributes, methods, and relations. They are used to recommend a ranked list of classes likely to contain defects for a system. We evaluated the efficiency of our approach according to three criteria: presence of defects, number of defects, and defect density in the top-ranked classes. We conducted experiments with small to large systems and made comparisons against well-known complexity and OO metrics. Results show that the design evolution metrics, when used in conjunction with those metrics, improve the identification of defective classes. In addition, they provide evidence that design evolution metrics make significantly better predictions of defect density than other metrics and, thus, can help in reducing the testing effort by focusing test activity on a reduced volume of code. (Abstract shortened by UMI.)