An efficient Draco lizard optimized stacked Bi-LSTM framework for risk mitigation and resource allocation in project...

Area of Science:

• Computer Science
• Artificial Intelligence
• Software Engineering

Background:

• Traditional project management struggles with dynamic risk mitigation and resource optimization.
• Subjective assessments and static tools are insufficient for complex, evolving projects.
• Advanced analytics, AI, and deep learning offer predictive insights for adaptive strategies.

Purpose of the Study:

• To propose advanced deep learning methods for improved risk reduction in software project management.
• To enhance resource allocation optimization within software project management systems.
• To develop a framework integrating stacked bidirectional long short-term networks (S-BiLSTM) and Draco lizard optimization (DLO).

Main Methods:

• Data preprocessing, including one-hot encoding and normalization, to handle irrelevance and reduce redundancy.
• Implementation of a stacked bidirectional long short-term network (S-BiLSTM) for joint risk prediction and resource scheduling.
• Application of the Draco lizard optimization (DLO) technique for optimal resource allocation based on project parameters.

Main Results:

• The proposed S-BiLSTM model achieved a predictive accuracy of 99.2% for project risk identification.
• Demonstrated significant improvements in workload distribution, resource optimization efficiency, and utilization rates.
• The DLO technique effectively optimized resource allocation considering factors like project complexity and team experience.

Conclusions:

• Advanced deep learning approaches, specifically S-BiLSTM and DLO, significantly enhance software project management.
• The integrated framework provides adaptive strategies for effective risk mitigation and resource optimization.
• The proposed method offers a data-driven solution outperforming traditional approaches in predictive accuracy and efficiency.