Benchmarking and Comparison of Software Project Human Resource Allocation Optimization Approaches

Abstract

For the Staffing and Scheduling a Software Project (SSSP), one has to find an allocation of resources to tasks while considering parameters such skills and availability to identify the optimal delivery of the project. Many approaches have been proposed that solve SSSP tasks by representing them as optimization problems and applying optimization techniques and heuristics. However, these approaches tend to vary in the parameters they consider, such as skill and availability, as well as the optimization techniques, which means their accuracy, performance, and applicability can vastly differ, making it difficult to select the most suitable approach for the problem at hand. The fundamental reason for this lack of comparative material lies in the absence of a systematic evaluation method that uses a validation dataset to benchmark SSSP approaches. We introduce an evaluation process for SSSP approaches together with benchmark data to address this problem. In addition, we present the initial evaluation of five SSSP approaches. The results shows that SSSP approaches solving identical challenges can differ in their computational time, preciseness of results and that our approach is capable of quantifying these differences. In addition, the results highlight that focused approaches generally outperform more sophisticated approaches for identical SSSP problems.

A Proposal of a simulation-optimization methodology for allocation of agencies with human resources on hexagonal tessellation

“…(Naveh et al, 2007) presented an analysis of the operation of IBM's personnel evaluation system, which uses descriptors, priorities, and constraint programming to solve the combinatorial problem while using a shortest route algorithm to reduce the distance between the employee residence and their workplace. In the allocation and scheduling of human resources for software development, (Al Khatib & Noppen, 2016) made a comparison between different approaches that mainly have used Genetic algorithms and Simulated Annealing. (Widianta et al, 2018) also made a comparison between easy-to-apply weight-based multicriteria decision methods for employee placement such as the Analytic Hierarchy Process, Simple additive Weighting, TOPSIS, and PROMETHEE.…”

Application project task scheduling using dolphin swarm technology

“…There are many challenges facing researchers in software project management issues, especially when conducting research on planning and scheduling, one of them is the lack of real data for the planning and scheduling phase of the project as the software companies reserve information related to employees, their salaries and their skills and do not provide any information on how to divide and distribute the work or any information related to management and leadership within the work team. Researchers have tried [24] to provide data For real projects in cooperation with a Jordanian software company, but that information is still few and limited, so researchers resort to using randomly generated data that simulate real projects when conducting research, and despite the good results achieved by the techniques and algorithms used, they are still not ready for use in tools and commercial applications [25].…”