World Modeler™ Product Features

• Secure identity transmission using access/refresh JWT
• Identity model including individual, organizational and system level principals
• Function-level role-based access control (RBAC)
• User/password authentication (including 2FA) or authenticate using external service
• Support for SSL/TLS client/server connections
• Log of all access requests

• Storage of multiple Decision Models in a hierarchical structure for easy organization and access to models
• Model ownership and access privileges controlled by RBAC
• Text-based repository search
• Template library that simplifies creation of new models based on existing patterns
• Storage of rich metadata for each model
• Archive system for deleted models, journaled for auditability

• Repository of references to external assets that are used in Decision Models. These include:
  • Other Decision Models
  • Libraries of functions
  • Data sources and queries
  • Machine learning models
  • Large Language Models (LLMs) and Agents
  • Web APIs to internal or external services

• Full self-service model management. No vendor involvement required to create, update, run, or integrate with decision models
• Functionality to create or import Decision Models
• Model template library that allows new models to be created from pre-configured, standardized baseline model frameworks.
• Ability to share models with other users with access and function-level permission control
• Easy interface to create and modify visual diagrams (Causal Decision Diagrams (CDDs), data flow, and more) and to link them dynamically to model elements
• Ability to view diagrammatic display and executable elements of the model simultaneously in a “split screen”
• Ability to create multiple diagrams for a single model, improving clarity and understandability.
• Ability to support iterative creation of models through iterative design cycles for models, through “sketch” designs, preliminary, and final stages.
• Functionality to apply metadata tags and expressions to model elements for advanced filtering and semantic analysis.
• Rich set of intelligent, “model aware” visualization features, including:
  • Color model diagram elements based on their role in the model
  • Functionality to filter or select elements by Tags or other data
  • UI elements that allow the user to extract and visualize individual causal chains
  • Functionality that automatically generates and lays out diagrams based on dependency chains
  • Display of model error conditions on diagrams
• Facility to annotate decision diagrams with markups (notes, shapes, etc.)
• Capacity to export decision diagrams to common graphical formats (PNG, JPG, SVG)
• Functionality to integrate model definition with Visual diagramming
• Automated model refactoring support by propagating changes to element names to all references
• Cyclic dependency analysis:
  • Analyze causal dependency graph for loops
  • Visually identify cyclic dependencies and show resolution status
  • Resolve cyclic dependencies with selected initializers

• Support for multiple versions of the same model
• Ability to preview or roll back a model to any earlier version
• Ability to fork a model at any version
• Functionality to detect differences between model versions
• Defined workflow states of models and allowed transitions between states
• Ability to lock models in production to protect them from inadvertent modification

• Ability to evaluate a model within diagramming environment
• Support for single-case (action + external set) evaluation or evaluate over a range of values
• Textual and visual representation of evaluation results
• Persistent record of all evaluation runs, which can be reviewed or exported at any time
• Integrated validation suite for testing model evaluations against specified validation conditions
• Support for simple data types (numbers, strings, dates, Boolean) and complex types (arrays, objects, functions)

• Batch Simulate Decision Models to:
  • Explore what-if scenarios
  • Determine sensitivity and risk
  • Find optimal scenarios
• Functionality to store simulation templates which can be run multiple times to explore the effects of different conditions
• User-definable input parameter generation including pre-defined options for:
  • Domain limits
  • Iterating over domain values
  • Sampling from domains
  • Limiting number of simulation runs by absolute limit and/or specifying stop conditions
  • Parameter generation guided by search for optimal values
• Support for parallelization simulation execution for scaling
• Ability to make all simulation results available for review and analysis either within World Modeler client software, or by exporting to third-party tools

• Optimize decision Actions and Interventions relative to a specified objective function.
• Many optimizers available, including:

  Derivative free

• • Downhill Simplex Method (Nelder-Mead)
  • Powell’s Direction Set Method

          Gradient methods

• • BFGS
  • L‑BFGS (Limited‑Memory BFGS)
  • Conjugate Gradient
  • Newton’s Method
  • Levenberg–Marquardt

          Evolutionary

• • CMA‑ES (Covariance Matrix Adaptation Evolution Strategy)
  • Differential Evolution
  • Particle Swarm Optimization

          Global

• • Simulated Annealing
• User-defined optimizers can be defined
• Time series optimization
• Parameterized and ensemble optimizations

• Ability to export model into OpenDI JSON format
• Extensive, secured REST API for enterprise integration or to support client applications. Almost all World Modeler functions are accessible from the API
• World Modeler Server™ API Programmers Resource Center

• Plugin interface that allows creation of custom add-on functionality. Add-ons can expose extensions to the REST API
• World Modeler functionality can be consumed in new or existing client applications using the Client Framework. This provides an easy-to-use wrapper for applications that use the World Modeler API.

• World Modeler Server is implemented as an ASP.NET Core Web API. World Modeler client applications are implemented in Angular as Web applications. These may be deployed using the following infrastructure configurations:
  • Cloud hosted
  • Private cloud
  • On-premise