Matrix-based project dataset parsers.

Area of Science:

• Project Management
• Data Science
• Algorithm Analysis

Background:

• Existing project datasets are fragmented, with heterogeneous file structures for simulated/real, individual/multiprojects, and single/multimodal attributes.
• This fragmentation limits algorithm testing to single datasets, hindering robust evaluation of scheduling and resource allocation methods.
• Diverse internal project structures make it challenging to ensure algorithm generalizability.

Purpose of the Study:

• To develop a unified parsing method for diverse project datasets.
• To enable comprehensive testing and comparison of scheduling and resource allocation algorithms.
• To support the modeling of structural flexibility in agile, hybrid, and extreme project management.

Main Methods:

• A novel parsing method designed to read multiple project types (simulated, real, individual, multiprojects, single/multimodal attributes).
• Incorporation of activity priorities and flexible dependencies to model structural adaptability.
• Development of a framework for building a large, integrated project database.

Main Results:

• The parsing method successfully integrates heterogeneous project data sources.
• Researchers can now utilize a unified database for testing algorithms across a wider range of project structures.
• The method facilitates the analysis of algorithms considering activity priorities and dependency flexibility.

Conclusions:

• The proposed parsing method enhances the ability to test and compare scheduling and resource allocation algorithms.
• It provides a foundation for developing more robust and generalizable project management algorithms.
• This approach supports advanced project management methodologies by accommodating structural flexibility.