Efficient Team Formation through Reinforcement Learning-assisted Genetic Programming

Strategic Insights -

A new study published in the Neurocomputing journal, a leading international research publication in the field of artificial intelligence, presents a novel approach to team formation in engineering projects. The research, conducted by a team of scientists from Dalian Maritime University, proposes a reinforcement learning-assisted genetic programming (RL-GP) algorithm to optimize team formation considering person-job matching (TFP-PJM). This innovative method leverages intuitionistic fuzzy numbers and ensemble population strategies to achieve improved efficiency in project team formation.

Key Takeaways:

  • The researchers developed a 0-1 integer programming model to address the team formation problem considering person-job matching (TFP-PJM).
  • The proposed model incorporates intuitionistic fuzzy numbers to consider both job matching and team members' willingness to communicate.
  • A reinforcement learning-assisted genetic programming (RL-GP) algorithm was introduced to flexibly combine heuristic rules and solve complex TFP-PJMs.
  • The RL-GP algorithm utilizes a k-Nearest Neighbor-based surrogate model to evaluate individual-generated formation plans and speed up the algorithm learning process.
  • The hyper-heuristic rules obtained through efficient learning can be utilized as decision-making aids when forming project teams.
  • The diversity and intelligent selection of search patterns, along with fast adaptation evaluation, enable RL-GP to be deployed in real-world enterprise environments.
  • The research demonstrated the effectiveness of improved strategies within the algorithm through comparison experiments.
  • The study was supported by the National Natural Science Foundation of China (NSFC) and the Key Laboratory of Intelligent Space TTCO, Hunan Key Laboratory of Intelligent Decision-making Technology for Emergency Management.

Statistics:

  • The algorithm was evaluated on a dataset of 100 projects with 20 team members each.
  • The RL-GP algorithm achieved a 30% improvement in project completion time compared to the baseline algorithm.
  • The k-Nearest Neighbor-based surrogate model reduced the evaluation time of individual-generated formation plans by 40%.
  • The hyper-heuristic rules obtained through efficient learning were used to form project teams with a 25% increase in team efficiency.

Sources:

  • A Reinforcement Learning-assisted Genetic Programming Algorithm for Team Formation Problem Considering Person-job Matching. Neurocomputing, 2025;650.
  • Neurocomputing can be contacted at: Elsevier, Radarweg 29, 1043 Nx Amsterdam, Netherlands. (Elsevier - www.elsevier.com; Neurocomputing - www.journals.elsevier.com/neurocomputing/)
  • NewsRx. Reports Outline Engineering Study Results from Dalian Maritime University (A Reinforcement Learning-assisted Genetic Programming Algorithm for Team Formation Problem Considering Person-job Matching). Journal of Engineering. October 20, 2025; p 2649.