Concepts and Methodologies for Modeling and Simulation

The objective of this chapter is to illustrate how agents can be used to facilitate development of next-generation simulators and assist in conducting simulation experiments. First, we introduce and define software agents and then characterize the three dimensions of the ADS framework that help explore both the use of agents for simulation and the use of simulation for agents. This is followed by the introduction of multisimulation and delineation of how agent-monitored and agent-assisted mechanisms facilitate its design and implementation. To underline the diversity of roles that agents can play in the overall M&S life cycle, we highlight the issues and challenges in goal-directed experimentation with simulation models and present an agent-assisted and model-driven experiment management strategy to effectively address these challenges.

The rapid development of high-performance modeling and simulation for complex systems (HPMSCS) is motivated by the application demands to support two types of users: (1) the high-end M&S users involved in complex systems of systems development and (2) users that require high-performance simulation cloud service on demand to utilize three types of simulation—mathematical, man-in-loop, and hardware-in-loop/embedded simulation. Those application demands raise great challenges to traditional M&S technology. This chapter examines these challenges and presents solution strategies developed by our research team.

Data plays an essential role in almost every aspect of computer modeling and simulation. Ören’s differentiation of the types of simulations and discussions of how data can impact simulations provide a conceptual framework to categorize the many existing works of using data in simulation. Existing works in the extant literature were developed from the perspective of modeling, e.g., using data to support model design, model calibration, and model validation. In this chapter, we focus on using data from the perspective of simulation, i.e., assimilating real-time sensor data into a running simulation model in the context of online simulation.

Epistemology is the branch of philosophy that deals with gaining knowledge. It is closely related to ontology, the branch that deals with questions such as “What is real?” and “What do we know?” When using modeling and simulation (M&S), we usually do so to either apply knowledge, in particular when we are using them for training and teaching, or to gain knowledge, for example, when doing analysis or conducting virtual experiments. But none of our models represents reality as it is. They are only valid within their limitations, which leads to the famous quote of Box that “all models are wrong.” The question is therefore: how can we learn something from these models? What are the epistemological foundations for us simulationists? To guide the reader on this path, we will start with an introduction to the philosophical fields of ontology and epistemology, leading to a short history of science, both from a simulationist view. These views shape our discussion on the use of models and resulting limits and constraints. The outcome of these analyses, as will be demonstrated, will lead to a grand challenge for the simulation community to evolve within the community of scientists by including epistemological perspectives in curricula for simulationists as a pillar of our profession.

Multimodeling approaches are increasingly required for simulating multifaceted systems across many scientific disciplines. Such approaches represent the system as a set of subsystem models, each with its own structure and behavior. Some multimodeling approaches use modeling methods to define how the subsystem structures and behaviors interact. However, modeling a system this way brings about subsystem and composition complexity that must be managed. The complexities of hybrid models resulting from the interactions of the composed models can be reduced using interaction models. Independently developing and utilizing such interaction models provides additional flexibility in system model design, modification, and execution for both the subsystem models and the resultant hybrid system model. This chapter discusses the use of the polyformalism model composition approach for researching human–environment dynamics with direct support for managing the complexity, which results from subsystem model interactions within this domain.

In this chapter, we present a formal, declarative, and visual model transformation methodology to map a domain conceptual model (CM) to a distributed simulation architecture model (DSAM). The approach adheres to the principles of model-driven engineering (MDE). A two-phased automatic transformation strategy is delineated to translate a field artillery conceptual model (ACM) into a high-level architecture (HLA) federation architecture model (FAM). The produced model is then compiled by the code generator to generate source code that can be executed on a distributed simulation runtime infrastructure. The presented mechanism is generic because the proposed abstract CM template can be extended and specialized into a domain-specific CM and transformed by adjusting the domain-specific components of the transformation rules. Generalizing from the ACM-to-FAM transformation case study, we propose a set of key design principles and an implementation framework as a step forward in achieving generic conceptual model (CM) transformations for publish/subscribe (P/S)-based distributed simulation infrastructures.

Multimodeling concepts allow designers of complex systems to organize their work better and to address the different components of a given system at the right level of abstraction. Agent-based modeling and simulation allows defining the behavior of the different components in a complex application with ease, allowing one to focus on the particular behavior of the different entities in the model. These concepts, pioneered by Prof. Ören, have had an important impact in the field of modeling and simulation. Here, we show how to use these ideas in combination with the DEVS formalism invented by Prof. Zeigler in order to build complex spatial models with focus on building construction and evacuation processes.

Different types of large-scale and complex modeling and simulation (M&S) applications are used in dozens of disciplines under diverse objectives such as acquisition, problem solving (analysis), education, entertainment, research, and training. Assessment of the quality of such M&S applications requires multifaceted knowledge and experience, and poses substantial technical and managerial challenges for researchers, practitioners, and engineers. The challenges can only be met by following a structured quality-centered approach throughout the entire M&S life cycle. This chapter presents dozens of quality indicators throughout the entire M&S life cycle. These quality indicators can be adopted to achieve success in a large-scale and complex M&S project.

Reproducible modeling and simulation research has been identified as one of the Modeling and Simulation (M&S) Grand Challenge activities. Recently, uncertainty quantification has seen a renewed emphasis. While methods for verification and validation (V&V) have been widely developed for discrete-event simulations, newer simulation approaches such as the agent-based, agent-directed, and multi-agent simulation approaches introduce new V&V challenges. The active elements in these newer approaches have greater heterogeneity, e.g., every agent can be unique, with complex attributes and behaviors. Those behaviors can result in actions based on interaction with other agents, the environment, and even the outcome of simulated or artificial intelligence. The simulation spaces are often less constrained, e.g., rather than a network of servers and queues, the space can be continuous 2D Euclidian space with multiple associated geographic information systems (GIS) layers influencing the behavior of the actors. Over the last decade, a multitude of techniques has been used in agent-based modeling and simulation (ABMS) to perform V&V as well as replication and reproducibility (R&R) of the models. In this chapter, we present an overview of contributions in V&V, quality assurance (QA), and R&R of simulation studies, with special focus on ABMS. We also discuss the lessons learnt in V&V and replication from a series of simulation experiments using agent-based models (ABMs).

Agent-based models are generative models to provide the unrecorded, yet important storylines to model users. Such generative models might have less accuracy in the prediction, but the provided insights would be invaluable even compared to the accurate predictions. In spite of this different value proposition, outsiders as well as insiders of the agent-based modeling community often concern about the accuracy of these generative models. Hence, this chapter intends to investigate the differences of the model for regeneration of the real world, i.e., agent-based model, and the model for accurate predictions of the real world, i.e., statistical model. The investigations include qualitative as well as quantitative comparisons of the two types of models. Qualitatively, the two types of models are contrasted by what they can contribute at different levels, i.e., conceptual contribution, design contribution, and prediction contribution. Quantitatively, an agent-based model for city commerce was compared to the corresponding statistical model. This particular comparison, again, casts light on the trade-off of different contributions from different models.

In this chapter, we present and demonstrate the advantages of the Generalized Discrete EVent system Specification (GDEVS) to build accurate discrete-event models of dynamic systems. These theoretical concepts are applied to the field of logic gate design and analysis in order to get more accurate and fast simulations. States are represented with linear piecewise trajectories contrary to the classical Boolean logic models where states have constant piecewise trajectories (0 and 1). With GDEVS models, the transition from a low level to a high one and vice versa is a linear trajectory and is more realistic than the instantaneous transitions of classical logic gate models. We also demonstrate that this accurate representation does not require any more computations than in Discrete EVent system Specification (DEVS).

Social science knowledge is best approached qualitatively, and efforts to force research into quantitative form (as with statistical analysis of, say, terrorist incidents versus alleged determinants) are often counterproductive because the data analysis of historical events is intellectually flawed, with problems of hidden variables, uncontrolled variables, and poor proxy variables. This chapter demonstrates that qualitative causal models can be more informative even if they cannot be used to make predictions as one would use a weather model or a model of whether precision-guided munitions would be able to destroy some particular target. They can be valuable for structuring issues, explaining what happens, and planning under uncertainty. Although the causal-loop or influence-diagram methods of system dynamics are powerful qualitatively, the uncertainties in dynamics are significantly worse even than those in a static situation. The chapter focuses on the static depiction and then discusses dynamics qualitatively.

While agent-based simulations have been a subject of a great deal of research in recent years, to date there is no framework for describing social agents that captures the uniqueness of human decision-making while remaining applicable across a wide variety of domains. We argue that if multiagent systems are to provide a proper computational model of both human decision-making and social interaction, then these structures and institutions and the cognitive capabilities of the agents that comprise them must be modeled to a level where computational complexity is not sacrificed on behalf of realism. This chapter examines the challenges facing frameworks describing cognitive agents embedded in a social environment. Following the analysis of these challenges, we introduce innovative mechanisms that allow agents to exhibit social behaviors by balancing their individual wants and needs with the concerns of the entire society while retaining a high level of cognition.