An Efficient Software Defect Prediction Model Using Optimized Tabu Search Branch and Bound Procedure

Software fault localization is considered to be one of the most tedious procedures that involves larger amount of time during the debugging of program. With this, there arises an increasing desire for software fault localization to be practiced with minimum amount of human intervention. This resulted with the design of several methods, each of which provides means to address the issues related to software fault localization to be more significant in its own notable and innovative manner. While the automatic structure based fault localization using genetic programming retains the program by avoiding a specific error, but failed to repair new types of bugs and programs. Most of the present software fault localization method overcomes the individual software failures and faults. However, an in-depth insight into the work reveals that, localization does not support several combinations of heuristics faults while performing software component testing. Finally, high automatic fault localization demand led to the proposal and development of software functionality on predicting the faults at an earlier stage with minimal prediction time. To overcome the defect on software fault localization, Tabu Search Fault Localization with Path Branch and Bound procedure on Software Engineering (TSFL-PBB) is proposed in this paper. TSFL-PBB divides the work into two phases. The first phase identifies (i.e.,) search doubtful software programming code which contain bugs (i.e.,) faults using Meta-heuristic Tabu search method. The mathematical operational based optimization checks with the immediate neighbors to handle different combinations of heuristics faults. The second phase of the TSFL-PBB software engineering model develops the Path Branch and Bound procedure. The branch and bound procedure in TSFL-PBB uses the travelling salesperson operation on localizing the faults at a faster prediction rate with higher readability and maintainability of software quality. Experiment is performed on factors such as fault prediction rate, processing time, repair cost.