A lecturer at the University of Human Development (UHD), Department of Computer Science, coauthored a scientific paper titled “A New Approach for Software Cost Estimation with a Hybrid Tabu Search and Invasive Weed Optimization Algorithms” and published in UHD Journal of Science and Technology. Dr Mazen Ismaeel Ghareb and two other authors propose a new approach to Software Cost Estimation by combining two algorithms, Tabu Search (TS) and Invasive Weed Optimization (IWO). Details can be found here on this link.

UHD Journal of Science and Technology (UHDJST) is a semi-annual peer-reviewed open-access journal published by UHD. It is indexed in some important databases. UHD publishes another journal called Journal of University of Human Development (JUHD). It covers humanity and social sciences topics, and it is too indexed in some important databases. Both journals are licensed by the Kurdistan Regional Government Ministry of Higher Education and Scientific Research which means that published articles can be used by authors for academic promotion (higher academic titles) at Kurdistan universities and higher education institutions. 

The paper’s abstract:

Due to the ever-increasing progress of software projects and their widespread impact on all industries, models must be designed and implemented to analyze and estimate costs and time. Until now, most of the software cost estimation (SCE) has been based on the analyst’s experiences and similar projects and these models are often inaccurate and inappropriate. The project will not be finished in the specified time and will include additional costs. Algorithmic models such as COCOMO are not very accurate in SCE. They are linear and the appropriate value for effort factors is not considered. On the other hand, artificial intelligence models have made significant progress in the cost estimation modeling of software projects in the past three decades. These models determine the correct value for effort factors through iteration and training, providing a more accurate estimate compared to algorithmic models. This paper employs a hybrid model incorporating the Tabu Search (TS) algorithm and the Invasive Weed Optimization (IWO) algorithm for SCE. IWO algorithm solutions are improved using the TS algorithm. The NASA60, NASA63, NASA93, KEMERER, and MAXWELL datasets are used for the evaluation. The proposed model has been able to reduce the MMRE rate compared to the IWO algorithm and the TS algorithm. The proposed model on the NASA60, NASA63, NASA93, KEMERER, and MAXWELL datasets obtained values of MMRE of 15.43, 17.05, 28.75, 58.43, and 22.46, respectively.