Optimizing software development requirements based on dependency relations

Software development planning and the optimization of requirements are increasingly complicated due to the complex technical and functional dependency relations among software development requirements. In this article, a directed graph model is adopted to describe the dependency relations among requirements from a global perspective. The concepts of value and cost were introduced into this model to describe resource limitations and employer expectations during the development of a software system. The resultant requirement sets are evaluated according to the ratio of the software value to its costs (i.e., value-to-cost ratio). Furthermore, we design a heuristic function which could consider one requirements’ value form the global view, and propose an A*-pruning algorithm based on the A* algorithm to obtain the requirement sequence set of a global optimum. The results generated with this approach are compared with those of five other methods under various goals and constraints. Experimental findings indicate that the A*-pruning algorithm can outperform other methods in terms of value-to-cost ratio. This work therefore provides an optimization scheme for requirement analysis in software engineering. © 2018, Ubiquitous International. All rights reserved.