[Paper Review] An Introduction to Classic DEVS

Yentl Van Tendeloo, Hans Vangheluwe|arXiv (Cornell University)|Jan 26, 2017

Simulation Techniques and Applications10 references3 citations

TL;DR

This paper introduces Classic DEVS, a formalism for modeling complex dynamic systems using discrete-event abstraction, emphasizing rigorous semantics, modularity, and hierarchical composition. It presents a bottom-up tutorial using a traffic light example, demonstrates simulation via an abstract simulator, and discusses applications in performance analysis and limitations addressed by variants like Parallel DEVS, Dynamic Structure DEVS, and Cell-DEVS.

ABSTRACT

DEVS is a popular formalism for modelling complex dynamic systems using a discrete-event abstraction. At this abstraction level, a timed sequence ofpertinent "events" input to a system (or internal, in the case of timeouts) cause instantaneous changes to the state of the system. Between events, the state does not change, resulting in a a piecewise constant state trajectory. Main advantages of DEVS are its rigorous formal definition, and its support for modular composition. This chapter introduces the Classic DEVS formalism in a bottom-up fashion, using a simple traffic light example. The syntax and operational semantics of Atomic (i.e., non-hierarchical) models are intruced first. The semantics of Coupled (hierarchical) models is then given by translation into Atomic DEVS models. As this formal "flattening" is not efficient, a modular abstract simulator which operates directly on the coupled model is also presented. This is the common basis for subsequent efficient implementations. We continue to actual applications of DEVS modelling and simulation, as seen in performance analysis for queueing systems. Finally, we present some of the shortcomings in the Classic DEVS formalism, and show solutions to them in the form of variants of the original formalism.

Motivation & Objective

• To provide a comprehensive, accessible introduction to Classic DEVS as a formalism for modeling complex dynamic systems using discrete-event abstraction.
• To demonstrate the formal syntax and operational semantics of atomic and coupled DEVS models through a running traffic light example.
• To present an abstract simulator that enables modular, efficient simulation of hierarchical DEVS models.
• To illustrate practical applications of DEVS in performance modeling, particularly for queueing systems.
• To identify key limitations of Classic DEVS and introduce major variants that address these shortcomings, such as dynamic structure and cellular modeling support.

Proposed method

• Introduces atomic DEVS models using a traffic light example, defining state sets and time advance functions for each state.
• Defines event-driven state transitions based on input events and internal timeouts, ensuring piecewise constant state trajectories.
• Presents coupled DEVS models via translation into atomic models, enabling hierarchical composition with ports and couplings.
• Develops an abstract simulator that operates directly on coupled models, improving efficiency over flat translation.
• Applies the formalism to a queueing system model to demonstrate utility in performance analysis and system evaluation.
• Proposes and explains key variants—Parallel DEVS, Dynamic Structure DEVS, and Cell-DEVS—each addressing specific limitations of Classic DEVS.

Experimental results

Research questions

• RQ1 How can a discrete-event formalism like Classic DEVS be used to model complex dynamic systems with rigorous semantics and modularity?
• RQ2 What are the core components and operational semantics of atomic and coupled DEVS models, and how do they support hierarchical system composition?
• RQ3 How does the abstract simulator enable efficient simulation of coupled DEVS models without full flattening?
• RQ4 In what ways do variants like Dynamic Structure DEVS and Cell-DEVS overcome limitations in structural flexibility and time granularity?
• RQ5 What are the practical applications of DEVS in performance modeling, particularly for queueing systems?

Key findings

• Classic DEVS provides a formally defined, modular framework for modeling discrete-event systems, with a clear separation between model and simulation kernel.
• The abstract simulator enables efficient simulation of coupled models by operating directly on the hierarchical structure, avoiding costly flattening.
• The traffic light example effectively illustrates the progression from atomic to coupled models, demonstrating event handling and state transitions.
• Applications to queueing systems show that DEVS is effective for performance analysis, particularly in modeling system throughput and waiting times.
• Variants such as Dynamic Structure DEVS and Cell-DEVS extend Classic DEVS to support dynamic structural changes and cellular modeling, respectively, enhancing applicability.
• Despite performance trade-offs, Parallel DEVS remains widely used due to its formal elegance and compatibility with existing tools and frameworks.