Application of swarm techniques to requirements tracing

Semantic Swarm Intelligence for Candidate Links

“…SI systems typically employ a relatively large population of agents, which interact both between themselves and the environment [4]. As a result, all these agents, being simple, unable to reason a decision just by themselves, share knowledge and optimize their behavior on the basis of the knowledge shared.…”

“…After an ant finds a food source, on its way back to the nest, it lays down a pheromone trail, which increases the possibility of the fellow ants locating the food source. The fellow ants then follow the pheromone trace [4] and if they locate the food source, on their way back to the nest they lay down pheromone trails as well. If they did not find the food source, or the pheromone trail is far from optimal and it takes the ants too much time to get back to the nest, the pheromone trails evaporate, giving an opportunity to the other ants to locate the same food source via shorter path.…”

“…Requirements tracing is a sub-area of requirements management within software engineering. It is mainly concerned with keeping track of the existence of requirements and providing bi-directional relationships between associated requirements [4]. The benefit for users is that the sub-area allows them to trace the origin of a requirement and keep track of every change made to it.…”

“…The candidate link evaluation assesses the previously generated links, in order to confirm that they are correct. At last, the traceability analysis deals with analyzing if there are use cases (lower-level) that satisfy a particular requirement (higher-level) [4]. In this work, we concentrate on adapting the Swarm Technique to the candidate link generation.…”

“…Candidate link generation is concerned with retrieving relevant information [4]. Swarm algorithm is used to rank retrieved low-level requirement elements that may be relevant to high-level requirements which is adapted for candidate link generation.…”