Technology
Synthetic Environments for Analysis and Simulation
Our simulations are built upon the SEAS (Synthetic Environments for Analysis and Simulation) technology. Originally designed to simulate a global industrial environment on a distributed architecture for use by commercial enterprises, SEAS has been expanded to factor in research from the fields of management, economics, and psychology. The result is a revolutionary new technology, blending easy-to-use visual interfaces and complex artificial intelligence: half simulation, half interactive war-game, and better than either alone. Senior decision makers have used SEAS to solve problems ranging from business strategies to disaster management.
How It Works
Using the SEAS technology, Simulex recreates the real world in its many aspects (business, economic, social, political, and infrastructure) using agent based modeling. Thousands of artificial agents represent the fine details of decision making (consumers, components, packages, personnel) and dozens of human players recreate the strategic aspects of interaction (countries, regulatory agencies, firms, channels, etc.). When calibrated with real data, this approach allows for depth as well as breadth of decision making to be faithfully represented. Simulation, gaming, experimentation, and decision support can then be brought together in an integrated fashion and firms can gain an unprecedented insight and edge on the competition.
For more information on how SEAS works and how firms use it, please see our Frequently Asked Questions section.
Why SEAS is Better
The SEAS technology is different from other simulation technologies in seven key ways:
- It allows for the incorporation of models from multiple domains (social, political, economic, business, epidemiological, etc.) as necessary.
- It allows for interactivity; i.e., human players in the loop.
- It scales all the way to tera-scale levels and may be deployed on systems ranging from off-the-shelf PCs to supercomputers.
- It links to existing enterprise systems so data may be extracted from existing databases.
- It is modular.
- Its configuration is totally transparent and is easily built and modified by the user without programming knowledge.
- It offers very easy-to-use Web-based interfaces to allow the human players to digest the hundreds of variables which are often tracked in real time. SEAS allows the player to seamlessly move between the real world and simulated world through common interfaces.
Using the SEAS technology, businesses can faithfully recreate their competitive landscape including customers, channels, competitors, and suppliers. A mix of artificial agents and humans in the loop provide strategic depth as well breadth of decision making. The environment developed for a particular industry may now be used to refine tactics, strategy, and training in any industry.
SimBridge
As the need to solve larger and more complex problems grows, simulation integration becomes paramount for the modeling and simulation community. SimBridge is a simulation integration engine that allows multiple simulations to interact more efficiently, as autonomous agents cooperating in a society. It is, in a sense, a virtual “bridge” created to specifically target heterogeneous simulations, so that a simulation representing human behavior can run autonomously to a physics-based simulation resulting in a faithful representation of the real world.
For example, in disaster management simulations, SimBridge allows three heterogeneous technologies—SEAS-VIS, JSAF, and SEAS-NRT—to work in concert. SEAS-VIS applies various economic, social, political, media, and networking models within a virtual world to model population behavior. JSAF models military actions, commuter and pedestrian movement, collateral damage, and the process of repairing a building within specific urban environments. SEAS-NRT links the behavior of individuals in SEAS-VIS with the entities and environment of JSAF, to model crowd behaviors within the context of a national or global simulation.
Each simulation is autonomously managed, meaning no non-scalable, centralized management mechanisms are needed to monitor the disparate components. Simulations and components are linked to information or to agreed-upon homogenous data types rather than to each other. Links between simulations are configured to emerge at run-time and adapt to dynamic changes, freeing the developers from coordinating their designs and removing the need for global mechanisms to regulate changes across all simulations. Simulations with diverse data formats, granularities, and semantic meanings of data interact through a process of ontology matching, granularity conversion, and syntax translation.