SimNexus The future of business simulation...

[ Home ] [ Software ] [ Strategic Partners ] [ Analysis Toolbox ] [ Contact Us ]

Primary Modeling Constructs

System Dynamics

System Dynamics modeling focuses on the interaction of components on the aggregate level, and simulates how quantities, called "levels" or "stocks", change over time.  Changes to the quantities in the model are defined by "flows" which represent increases and decreases to the quantities within the "stocks".  These "stocks" and "flows" are further defined by causal relationships, or "links", that define relationships between objects in the model allowing for the simulation of feedback and delays.   The modeling structure uses these stocks, flows, and causal relationships to represent partial differential equations in an intuitive graphical user interface.  System Dynamics models are traditionally run in a continuous time series. 

Agent Modeling

Agent Modeling focuses on the interaction of components at the molecular level, and can simulate a wide array of complex interactions.  In advanced implementations of Agent Modeling the agents are represented as interactive players that can influence and be influenced by other factors within their environment.  Agents can be created with an object oriented structure that allows for complex aggregations, while their individual nature allows for cascading and recombinant interactions.  The also have the ability to modify their behavior based on a variety of artificial intelligence "learning" algorithms.  Agent Modeling supports both continuous and discrete event modeling in an integrated environment.

Software Development Tools

AnyLogic

AnyLogic is a graphical user interface that allows for the creation of Java based simulations.  It provides an intuitive graphical representation of an object oriented modeling framework that can incorporate both Agent Modeling and System Dynamics structures.  In addition to providing base classes for development in our two primary modeling constructs, it can also incorporate other existing Java code as well as custom Java code.  Because the end product is Java code, the tool is extremely extensible and can be run on any platform.

Java

Java has emerged as the language of choice for the development of complex software simulations.  The ubiquitous nature of Java and its inherent portability make it an obvious choice as a development tool.  Many of the algorithms that we use in our models are written in Java for incorporation with the AnyLogic modeling framework.

Graphical Environment

3D Active Chart

We have teamed with the makers of 3D Active Chart to provide a three dimensional data visualization environment.  This allows the data review effort to be more experiential, and is seamlessly integrated into our simulations.

Advanced Algorithms

Searches and Optimization

BREADTH FIRST

DEPTH FIRST

HEURISTIC SEARCH

GENETIC ALGORITHMS

HILL CLIMB

SIMULATED ANNEALING

Learning Systems

BAYESIAN NETWORKS

NEURAL NETWORKS

DECISION TREES

Reasoning Systems

FORWARD AND BACKWARD CHAINING

FUZZY LOGIC

Utility Functions

DISCRETE CHOICE LOGIT FUNCTIONS

CONJOINT ANALYSIS

Customer Relationship Modeling

MARKOV CHAIN MODELS

Portfolio Analysis

PORTFOLIO OPTIMIZATION

EFFICIENT FRONTIER

REAL OPTIONS

DERIVATIVES