Francisco Javier Orellana

Ant Colony Optimization for the Next Release Problem: A Comparative Study

The selection of the enhancements to be included in the next software release is a complex task in every software development. Customers demand their own software enhancements, but all of them cannot be included in the software product, mainly due to the existence limited resources. In most of the cases, it is not feasible to develop all the new functionalities suggested by customers. Hence each new feature competes against each other to be included in the next release. This problem of minimizing development effort and maximizing customers’ satisfaction is known as the next release problem (NRP). In this work we study the NRP problem as an optimisation problem. We use and describe three different meta-heuristic search techniques for solving NRP: simulated annealing, genetic algorithms and ant colony system (specifically, we show how to adapt the ant colony system to NRP). All of them obtain good but possibly sub optimal solution. Also we make a comparative study of these techniques on a case study. Furthermore, we have observed that the sub optimal solutions found applying these techniques include a high percentage of the requirements considered as most important by each individual customer.

Multi-objective ant colony optimization for requirements selection

The selection of a set of requirements between all the requirements previously defined by customers is an important process, repeated at the beginning of each development step when an incremental or agile software development approach is adopted. The set of selected requirements will be developed during the actual iteration. This selection problem can be reformulated as a search problem, allowing its treatment with metaheuristic optimization techniques. This paper studies how to apply Ant Colony Optimization algorithms to select requirements. First, we describe this problem formally extending an earlier version of the problem, and introduce a method based on Ant Colony System to find a variety of efficient solutions. The performance achieved by the Ant Colony System is compared with that of Greedy Randomized Adaptive Search Procedure and Non-dominated Sorting Genetic Algorithm, by means of computational experiments carried out on two instances of the problem constructed from data provided by the experts.

Requirements interaction in the next release problem

The selection of a set of requirements between all those proposed by the customers is an important process in software development, that can be addressed using heuristic optimization techniques. Dependencies or interactions between requirements can be defined to denote common situations in software development: requirements that follow an order of precedence, requiments exclusive of each other, requirements that must be included at the same time, etc. This paper shows how requirements interactions affect the search space explored by optimization algorithms. Three search techniques, i.e. a greedy randomized adaptive search procedure (GRASP), a genetic algorithm (GA) and an ant colony system (ACS), have been adapted to the requirements selection problem considering interaction between requirements. We describe the adaptation of the three meta-heuristic algorithms to solve this problem and compare their performance.

Architecture for the use of synergies between knowledge engineering and requirements engineering

The application of Artificial Intelligence techniques in the processes of Software Engineering is achieving good results in those activities that require the use of expert knowledge. Within Software Engineering, the activities related to requirements become a suitable target for these techniques, since a good or bad execution of these tasks has a strong impact in the quality of the final software product. Hence, a tool to support the decision makers during these activities is highly desired. This work presents a three-layer architecture, which provides a seamless integration between Knowledge Engineering and Requirement Engineering. The architecture is instantiated into a CARE (Computer-Aided Engineering Requirement) tool that integrates some Artificial Intelligence techniques: Requisites, a Bayesian network used to validate the specification of the requirements of a project, and metaheuristic techniques (simulated annealing, genetic algorithm and an ant colony system) to the selection of the requirements that have to be included into the final software product.

Metaheuristic Aided Software Features Assembly

One significant task addressed during software development project is to determine which features should be covered by the application that is being developed. This problem is known as the Next Release Problem (NRP) and has been solved using metaheuristic search techniques. We show how to apply these techniques by its embedding into a requirement management tool as an assistant functionality. We have called this new utility MASA (Metaheuristic Aided Software Features Assembly).