Adelinde M. Uhrmacher

Multi-Agent Systems: Simulation and Applications

Methodological Guidelines for Modeling and Developing MAS-Based Simulations The intersection of agents, modeling, simulation, and application domains has been the subject of active research for over two decades. Although agents and simulation have been used effectively in a variety of application domains, much of the supporting research remains scattered in the literature, too often leaving scientists to develop multi-agent system (MAS) models and simulations from scratch. Multi-Agent Systems: Simulation and Applications provides an overdue review of the wide ranging facets of MAS simulation, including methodological and application-oriented guidelines. This comprehensive resource reviews two decades of research in the intersection of MAS, simulation, and different application domains. It provides scientists and developers with disciplined engineering approaches to modeling and developing MAS-based simulations. After providing an overview of the fields history and its basic principles, as well as cataloging the various simulation engines for MAS, the book devotes three sections to current and emerging approaches and applications. Simulation for MAS explains simulation support for agent decision making, the use of simulation for the design of self-organizing systems, the role of software architecture in simulating MAS, and the use of simulation for studying learning and stigmergic interaction. MAS for Simulation discusses an agent-based framework for symbiotic simulation, the use of country databases and expert systems for agent-based modeling of social systems, crowd-behavior modeling, agent-based modeling and simulation of adult stem cells, and agents for traffic simulation. Tools presents a number of representative platforms and tools for MAS and simulation, including Jason, James II, SeSAm, and RoboCup Rescue. Complete with over 200 figures and formulas, this reference book provides the necessary overview of experiences with MAS simulation and the tools needed to exploit simulation in MAS for future research in a vast array of applications including home security, computational systems biology, and traffic management.

Dynamic structures in modeling and simulation: a reflective approach

As the number of flexible, adaptable systems grows so does the need for specification and analysis tools that support adaptable system structures. The increasing number of simulation tools that equip models with the capability of changing their behavior patterns, composition, and interactions raises the desire for a theoretical and methodological approach. A formalism is introduced based on DEVS which emphasizes the reflective nature of variable structure models. The proposed formalism and DEVS are shown to be bisimilar, which emphasizes the role of variable structure models as an agency of modularization. The formalism is used to reveal general problems and solutions in implementing variable structure models.

Rule-based multi-level modeling of cell biological systems

BackgroundProteins, individual cells, and cell populations denote different levels of an organizational hierarchy, each of which with its own dynamics. Multi-level modeling is concerned with describing a system at these different levels and relating their dynamics. Rule-based modeling has increasingly attracted attention due to enabling a concise and compact description of biochemical systems. In addition, it allows different methods for model analysis, since more than one semantics can be defined for the same syntax.ResultsMulti-level modeling implies the hierarchical nesting of model entities and explicit support for downward and upward causation between different levels. Concepts to support multi-level modeling in a rule-based language are identified. To those belong rule schemata, hierarchical nesting of species, assigning attributes and solutions to species at each level and preserving content of nested species while applying rules. Further necessities are the ability to apply rules and flexibly define reaction rate kinetics and constraints on nested species as well as species that are nested within others. An example model is presented that analyses the interplay of an intracellular control circuit with states at cell level, its relation to cell division, and connections to intercellular communication within a population of cells. The example is described in ML-Rules - a rule-based multi-level approach that has been realized within the plug-in-based modeling and simulation framework JAMES II.ConclusionsRule-based languages are a suitable starting point for developing a concise and compact language for multi-level modeling of cell biological systems. The combination of nesting species, assigning attributes, and constraining reactions according to these attributes is crucial in achieving the desired expressiveness. Rule schemata allow a concise and compact description of complex models. As a result, the presented approach facilitates developing and maintaining multi-level models that, for instance, interrelate intracellular and intercellular dynamics.

Modeling dynamic environments in multi-agent simulation

Real environments in which agents operate are inherently dynamic—the environment changes beyond the agents’ control. We advocate that, for multi-agent simulation, this dynamism must be modeled explicitly as part of the simulated environment, preferably using concepts and constructs that relate to the real world. In this paper, we describe such concepts and constructs, and we provide a formal framework to unambiguously specify their relations and meaning. We apply the formal framework to model a dynamic RoboCup Soccer environment and elaborate on how the framework poses new challenges for exploring the modeling of environments in multi-agent simulation.

A Spatial Extension to the π Calculus

Spatial dynamics receive increasing attention in Systems Biology and require suitable modeling and simulation approaches. So far, modeling formalisms have focused on population-based approaches or place and move individuals relative to each other in space. SpacePi extends the @p calculus by time and space. @p processes are embedded into a vector space and move individually. Only processes that are sufficiently close can communicate. The operational semantics of SpacePi defines the interplay between movement, communication, and time-triggered events. A model describing the phototaxis of the Euglena micro-organism is presented as a practical example. The formalisms use and generality is discussed with respect to the modeling of molecular biological processes like diffusion, active transportation in cell signaling, and spatial structures.

Distributed, parallel simulation of multiple, deliberative agents

Multi agent systems comprise multiple, deliberative agents embedded in and recreating patterns of interactions. Each agents execution consumes considerable storage and calculation capacities. For testing multi agent systems, distributed parallel simulation techniques are required that take the dynamic pattern of composition and interaction of multi-agent systems into account. Analyzing the behavior of agents in virtual, dynamic environments necessitates relating the simulation time to the actual execution time of agents. Since the execution time of deliberative components can hardly be foretold, conservative techniques based on lookahead are not applicable. On the other hand, optimistic techniques become very expensive if mobile agents and the creation and deletion of model components are affected by a rollback. The developed simulation layer of JAMES (a Java Based Agent Modeling Environment for Simulation) implements a moderately optimistic strategy which splits simulation and external deliberation into different threads and allows simulation and deliberation to proceed concurrently by utilizing simulation events as synchronization points.

Modeling and simulation of mobile agents

Abstract Agent-oriented software implies the realization of software components, which are mobile, autonomous, and solve problems by creating new software components during run-time, moving between locations, initiating or joining groups of other software components. Modeling and simulating multiagent systems requires specific mechanisms for variable structure modeling. JAMES, a Java-Based Agent Modeling Environment for Simulation, realizes variable structure models including mobility from the perspective of single autonomous agents. JAMES itself is based on parallel DEVS and adopts its abstract simulator model. Simulation takes place as a sending of messages between concurrently active and locally distributed entities which reflect the model’s current structure. Thus, modeling and simulation are coined equally by an agent-based perspective.

Combining micro and macro-modeling in DEVS for computational biology

In computational biology there is an increasing need to combine micro and macro views of the system of interest. Therefore, explicit means to describe micro and macro level and the downward and upward causation that link both are required. Multi-Level-DEVS (or m^-DEVS) supports an explicit description of macro and micro level, information at macro level can be accessed from micro level and vice versa, micro models can be synchronously activated by the macro model and also the micro models can trigger the dynamics at macro level. To link both levels, different methods are combined, to those belong, value coupling, synchronous activations, variable ports, and invariants. The execution semantic of the formalism is given by an abstract simulator and its use is illustrated based on an small extract of the Wnt pathway.

Discrete event multi-level models for systems biology

Diverse modeling and simulation methods are being applied in the area of Systems Biology. Most models in Systems Biology can easily be located within the space that is spanned by three dimensions of modeling: continuous and discrete; quantitative and qualitative; stochastic and deterministic. These dimensions are not entirely independent nor are they exclusive. Many modeling approaches are hybrid as they combine continuous and discrete, quantitative and qualitative, stochastic and deterministic aspects. Another important aspect for the distinction of modeling approaches is at which level a model describes a system: is it at the “macro” level, at the “micro” level, or at multiple levels of organization. Although multi-level models can be located anywhere in the space spanned by the three dimensions of modeling and simulation, clustering tendencies can be observed whose implications are discussed and illustrated by moving from a continuous, deterministic quantitative macro model to a stochastic discrete-event semi-quantitative multi-level model.

Planning agents in JAMES

Testing is an obligatory step in developing multiagent systems. For testing multiagent systems in virtual, dynamic environments, simulation systems are required that support a modular, declarative construction of experimental frames, that facilitate the embedding of a variety of agent architectures and that allow an efficient parallel, distributed execution. We introduce the system JAMES (a Java based agent modeling environment for simulation). In JAMES, agents and their dynamic environment are modeled as reflective, time-triggered state automata. Its possibilities to compose experimental frames based on predefined components, to express temporal interdependencies, to capture the phenomenon of proactiveness and reflectivity of agents are illuminated by experiments with planning agents. The underlying planning system is a general-purpose system, about which no empirical results exist besides traditional static benchmark tests. We analyze the interplay between heuristics for selecting goals, viewing range, commitment strategies, explorativeness, and trust in the persistence of the world and uncover properties of the the agent, the planning engine, and the chosen test scenario: TILEWORLD.