Combining particle swarm optimization and genetic algorithms to improve software effort estimation

Ehsan Nasr, Keyvan Mohebbi

Abstract


Analogy-based estimation compares a project to finished ones to estimate software development work. Despite its numerous advantages, this strategy doesn't work well when project features have varying priorities or dependencies. To fix the problem, optimization techniques, especially meta-heuristic algorithms are often used, however, they might get stuck in the local optimum. This research combines particle swarm and genetics techniques to optimize feature weight globally. This hybrid method increases the likelihood of discovering the global optimal and solving the local optimal problem by employing particle motion in state space, composition, and mutation. Using this procedure, the needed weights for the project characteristics are derived and utilized to estimate the work. The suggested method was tested using Maxwell and Desharnais datasets. The experiment improved the mean magnitude of relative error (MMRE), the median magnitude of relative error (MdMRE), and prediction (PRED).

Full Text:

PDF


DOI: http://doi.org/10.11591/ijaas.v11.i3.pp199-210

Refbacks

  • There are currently no refbacks.


International Journal of Advances in Applied Sciences (IJAAS)
p-ISSN 2252-8814, e-ISSN 2722-2594

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.


Web Analytics View IJAAS Stats