A Multimodel Framework for the Analysis of Processes Described Using UML

UML activity diagrams are widely used for process representation, but their lack of formal semantics limits rigorous analysis.This paper presents a multimodel framework that translates UML activity diagrams into timed Petri net models for validation and simulation.The approach enables the detection of structural and behavioral issues such as deadlocks, unreachable states, and performance bottlenecks.To illustrate the methodology, we apply it to two optimization algorithms -Combinatorial Artificial Bee Colony (CABC) and Discrete Chicken Swarm Optimization (DCSO) -showing how their UML-based descriptions can be systematically transformed and analyzed.The proposed framework improves model precision, supports iterative refinement, and provides a formal basis for evaluating dynamic, concurrent processes.