Process-Centric Digital Twin for Agile Software Development Resource Allocation: A What-If Simulation Approach for Decision Support
Abstract
Resource allocation in Agile software development is often performed reactively, limiting the ability of project managers to evaluate the impact of changing requirements, workload fluctuations, resource availability, and task dependencies before decisions are implemented. Existing project management tools provide visibility into project progress but offer limited support for simulation-driven evaluation of alternative resource allocation strategies. This paper proposes a Process-Centric Digital Twin (PC-DT) framework that integrates workflow discovery and what-if simulation to support proactive resource allocation planning in Agile environments. Using the Design Science Research methodology, operational workflows and resource allocation scenarios were identified through document analysis, stakeholder interviews, and direct observation within a higher education IT center. The discovered workflow was formalized into a PC-DT model consisting of Sprint, Task, Resource, Role, and Skill entities, together with workload and dependency relationships. Based on this model, a simulation engine was developed to evaluate eight what-if scenarios, including change requests, team scaling, task redistribution, backlog reprioritization, partial resource availability, and dependency cascade analysis. A web-based prototype of the framework was developed and evaluated through scenario-based simulations that utilized historical project data alongside structured expert judgment. Results show that different resource allocation strategies produce distinct impacts on sprint velocity, completion rate, resource utilization, and project cost. Expert evaluation indicated strong model fidelity (4.2/5) and decision-support utility (4.4/5). This study contributes a workflow discovery approach, a formal PC-DT model for Agile software development, and a simulation-driven decision-support mechanism for resource allocation planning.