It featured many updates aimed at simplifying model building, including support for multimethod modeling, a decreased need for coding, renewed libraries, and other usability improvements. The new version was focused on business simulation in different industries. The first version of AnyLogic was V4, because the numbering continues the numbering of COVERS 3.0.ĪnyLogic 5 was released in 2003. and any combination of these approaches within a single model. The tool was named AnyLogic, because it supported all three well-known modeling approaches: system dynamics, discrete event simulation, Agent-based modeling. The resulting software was released in 2000 and featured the latest information technologies: an object-oriented approach, elements of the UML standard, the use of Java, and a modern GUI. Development emphasis was placed on applied methods: simulation, performance analysis, behavior of stochastic systems, optimization and visualization. In 1998, the success of this research inspired the DCN laboratory to organize a company with the mission of developing a new generation of simulation software. The tool was developed with the help of a research grant from Hewlett-Packard (Commonly known as HP). This system allowed graphical modeling notation to be used for describing system structure and behavior.
The Distributed Computer Network (DCN) research group at Saint Petersburg Polytechnic University developed a software system for the analysis of program correctness the new tool was named COVERS (Concurrent Verification and Simulation). This approach was applied to the analysis of correctness of parallel and distributed programs. In the early 1990s, there was a big interest in the mathematical approach to modeling and simulation of parallel processes.
It is considered to be among the major players in the simulation industry, especially within the domain of business processes is acknowledged to be a powerful tool. ĪnyLogic is used to simulate: markets and competition, healthcare, manufacturing, supply chains and logistics, retail, business processes, social and ecosystem dynamics, defense, project and asset management, pedestrian dynamics and road traffic, IT, and aerospace. AnyLogic is cross-platform simulation software that works on Windows, macOS and Linux. It supports agent-based, discrete event, and system dynamics simulation methodologies.
policy.English, Portuguese, Russian, German, Chinese, SpanishĪnyLogic is a multimethod simulation modeling tool developed by The AnyLogic Company (formerly XJ Technologies). # NOTE: This is required to be called for correct checkpoint saving by ALPypeRL. Print( f"Checkpoint saved in directory ' '") 'exported_model_loc': './resources/exported_models/cartpole_v0', Num_consecutive_worker_failures_tolerance = 3 checkpoint_dir = "./resources/trained_policies/cartpole_v0" # Initialize and configure policy using `rllib`. ppo import PPOConfig # Set checkpoint directory. trained policy deployment and evaluation).įrom alpyperl import AnyLogicEnv from ray. Because of that, ALPypeRL has certain dependencies to it (e.g. Ray rllib is an industry leading open source package for Reinforcement Learning. NOTE: ALPypeRL has been developed using ray rllib as the base RL framework. No licence is required for single instance experiments. There is a more comprehensive documentation available that includes numerous examples to help you understand the basic functionalities in greater detail. Leverage on the AnyLogic rich visualization while training or evaluating.Identify and replicate failed runs by having control on the seed used for each run.Debug your AnyLogic models during training ( this is a special feature unique to ALPypeRL that improves the user experience during model debugging remarkably).Deploy and evaluate your trained policy from AnyLogic.Scale your training by launching many AnyLogic models simultaneously ( requires an exported model).Connect your AnyLogic model to a reinforcement learning framework of your choice (e.g.ALPypeRL or AnyLogic Python Pipe for Reinforcement Learning is an open source library for connecting AnyLogic simulation models with reinforcement learning frameworks that are compatible with OpenAI Gymnasium interface (single agent).