AnyLogic

AnyLogic is a multi-method simulation tool that is developed by The AnyLogic Company.

The story of AnyLogic

Early 1990s showed you a great interest in a mathematical approach to the modeling and simulation of parallel processes. This approach could be applied to analyze the accuracy of parallel and distributed programs. The " Distributed Network Research Group" (Distributed Computer Network, DCN ) of the Technical University of St Petersburg has developed a software system for the analysis of the accuracy of the programs; the new tool was COVERS (Concurrent Verification and Simulation) called. This system allowed a graphical model representation for system structure and behavior. The tool was developed in the research for Hewlett Packard.

In 1998, inspired by the success of this research, the DCN- research institute founded a company with the aim to develop a modern simulation software. In developing the methods used were stressed: simulation, performance analysis, behavioral and stochastic systems, organization and visualization. Published in 2000, new software had the latest developments in information technology as the basis: object-oriented approach, elements of the standard UML, modern Java programming language and graphical user interface.

The tool has been AnyLogic called because it supports all three known modeling approaches:

• System Dynamics (SD )
• Event simulation,
• Agent-based modeling.

Any combination of these approaches is possible within a single model. The first version of AnyLogic AnyLogic is 4, because the numbering of COVERS 3.0 was continued. In 2003 AnyLogic 5 has been released. The focus was placed on business simulation in the following applications:

• Market and competition,
• Healthcare,
• Production and manufacturing,
• Supply chains,
• Logistics,
• Retail,
• Business processes,
• Social and ecosystem dynamics,
• Defense,
• Project and resource planning,
• IT infrastructure,
• Pedestrian flow and pedestrian dynamics,
• Aviation industry,
• Photovoltaics.

The latest version, AnyLogic 6 was published in 2007. The platform for the AnyLogic 6 model development environment is Eclipse. AnyLogic 6 is a cross-platform simulation software, especially since it can be run on Windows, Mac OS and Linux.

AnyLogic and Java

AnyLogic includes a graphical modeling language and also allows the user to extend simulation models to Java code. The Java character of AnyLogic is suitable both for individual model extensions via Java coding, as well as for the design of Java applets that can be opened in any standard browser. These applets make AnyLogic models easily shippable and loaded on websites. In addition to the Java applets, the Professional version allows the design and distribution of Java run- time applications to users. This exclusive Java application can serve as a basis for decision support tools.

Multi-method simulation modeling

AnyLogic models can on any major simulation modeling paradigms based: event oriented or process-centric (DE), system dynamics ( SD), and agent-based (AB).

The system dynamics and the discrete event simulation approaches are traditional; the agent-based approach is a modern method. Technically, the system dynamics approach deals mainly with continuous processes, the "event - discrete " ( meant are all descendants of GPSS, also known as process- centered approach to simulation ) and the agent-based models go against mostly work in discrete time, ie they change from one event to other.

System dynamics and discrete event simulation has been historically taught at universities various student groups, and btw Management, Economics and Operations Research - engineers. Thus, there are two different Sections that do not communicate with each other.

Agent-based modeling was until recently almost exclusively an academic school subject. Nevertheless, caused the increasing request for a global Unternehmensoptimisierung that leading modelers did consider, approaches to combine, thus giving them a better insight into the complex, diverse, and interdependent processes.

How correspond modeling approaches the levels of abstraction? The System Dynamics, which deals with aggregates is, of course, applied to the highest level of abstraction. Event-oriented modeling is used at low to medium abstraction. As for the agent-based modeling, this technique is applied to all levels of abstraction, also allows an agent to objects of different type and size of model: on "physical " level, agents can, for example, as pedestrians, vehicles, robots, mid-level - as a customer, be designed as a competitor - at the highest level.

AnyLogic allows the modeler this simulation approaches to link within the same model. There is no fixed hierarchy. Thus, one could for example make a model for the transport package in the industry, where the charge carriers are modeled as agents that individually act / react, while the internal activities of their transport and infrastructure networks could be modeled by discrete event simulation. Similarly, one can model consumers as agents whose entire practice a system dynamic model powered by detecting processes such as income and expenditure, which must not be tied to individual agents. This linguistically mixed approach is directly applicable to a variety of complex Modellierunsproblemen that can be modeled by jedwelchen an approach, albeit with compromises.

Simulation language

AnyLogic simulation language consists of the following elements:

• Stock & Flow charts are used for system dynamics modeling.
• State diagrams are mainly used in the agent-based modeling for the determination of agent behavior. They are also used in the event-oriented modeling, for example, to simulate Maschinausfällen.
• Functional diagrams are used to determine algorithms. They can be used in the event-oriented modeling, for example, call distribution, or the agent-based modeling, for example, in the arbitration logic of the agent.
• Process flow diagrams illustrate the basic design in determining the processes in event-oriented modeling dar. Looking at this flowchart, it is apparent why the event-based method is often mentioned as a process- centered.

The language includes: modeling structures at low level (variables, equations, parameters, events, etc. ), design shapes (lines, polylines, ellipses ), analysis capabilities ( data, histograms, recording points), compound tools, standard images and test framework development.

AnyLogic libraries

AnyLogic includes the following standard libraries:

• The Enterprise Library is designed so that it supports DE simulation in areas such as production, supply chains, logistics and healthcare. Using the Enterprise Library objects you can model real systems, in the sense of records (transactions, customers, products, components, vehicles, etc.) processes ( sequences of operations that normally lines, delays, resource utilization include) and resources. The processes are specified in the form of flow charts.
• The Pedestrian Library is suitable for the simulation of passenger flows in a "physical" environment. It allows the design of building models with intensive passersby transport (such as subway stations, security checkpoints, etc.) or road models ( large number of passers-by ). The models support statistical data collection on pedestrians density in different zones. This ensures an acceptable performance of departments with a hypothetical stress, estimated from the dwell times in specific areas and detects potential problems related to the internal geometry - such as the impact of too many obstacles - and other applications. In the, with the Pedestrian Library designed models, passers move in continuous space, react to various obstacles (walls, various areas) and to other passers-by. Passers-by are simulated as interacting agents with complex behaviors, but the AnyLogic Pedestrian Library provides a high level interface for rapid creation of passers models in the form of flow charts.
• The Railyard Library supports modeling, simulation, and visualization of processes within a station from anybody complexity or size. The station models can be associated with discrete event or agent based models in terms of: loading and unloading, distribution of resources, maintenance, business processes and other transportation activities.

Apart from these standard libraries, users can design and distribute their own libraries.