Smartalloc: A model based on Machine Learning for human resource allocation in projects

• Eduardo J. Kieling Unisinos
• Felipe C. Rodrigues Unisinos
• Alexsandro Filippetto Unisinos
• Jorge L. V. Barbosa Unisinos

Resumo

This article presents the Smartalloc model for human resource allocations in projects based on machine learning. The model learns about the allocation strategies used by the organization over time and makes recommendations based on this information. The model has two scientific contributions, based on the study of related works: (1) allows the choice of the strategic objective of the organization (cost, time or quality) in the definition of the resource allocation strategy; (2) uses the historical allocations of previous projects. A prototype was implemented and applied in an evaluation involving 2 project managers of 2 organizations who answered structured research in the Technology Acceptance Model (TAM) methodology, confirming the usability of Smartalloc. Then, the Accuracy calculation of the machine learning algorithm was measured, whose ideal value should be 1. In 6 projects in the first company, the average was 0.77. In the second company, the average was 0.70 in 3 projects. Both project managers considered the Smartalloc model to be useful in allocating human resources to projects.