Eco Modeling System: Easy Simulations for Spatial Big Data

Description:

Eco Modeling System: Easy Simulations for Spatial “Big Data”

Docket # 14-3036

Agent-based models (ABMs) are computerized simulations of the behavior of many separate individuals, leading to predictions about the system as a whole. ABMs are used in academia and industry to explain emergent phenomena and to predict outcomes in untested scenarios. There are considerable barriers to entry for potential ABM users. Commercial ABM software, used to simulate industrial designs and urban plans, can cost more than $10,000 per license. Free software used in academia can be complex and require considerable computer programming experience.

Researchers at the Lewis-Sigler Institute for Integrative Genomics at Princeton University have developed a novel ABM platform, called “Eco,” that facilitates the rapid development of fast, spatially explicit ABMs. Their system utilizes a novel computer language whose grammar and syntax map onto familiar visual cues such as flow charts, and whose execution is automatically optimized by the computer prior to execution. The result is a system that can be used by subject area experts to develop powerful models in a variety of research areas, without the need for technology consultants or in-house programming expertise. Major application areas include epidemiology, political forecasting, design and marketing.

Applications:

· Creation of ABMs without writing code

· Distillation of “big data” for predictive purposes

· Development of powerful models in

o Epidemiology

o Political forecasting

o Design

o Marketing

Advantages:

· Grammar and syntax map onto familiar visual cues such as flow charts

· Execution is automatically optimized

· Fast

· Low cost

· Can be deployed as an interactive, “cloud-based” service

Background

Eco is a compiled computer platform for the easy development of spatially explicit models. Originally created to investigate community dynamics and social behavior within microbial ecosystems, Eco enables subject area experts to simulate research hypotheses with minimal training and lead time. A compiled computer language is a language in which the entire program is translated into computer-readable “byte code” before execution. The original, human-readable instructions are then ignored. Compiled languages can make intelligent choices about program execution, potentially resulting in large speed increases, error checking and syntactic simplicity.

Existing platforms in widespread use are either interpreted languages (e.g., NetLogo), “libraries” of computer code for use with general-purpose programming languages (e.g., MASON, Swarm, Repast), or hybrid systems that involve both high-level components and programming in general-purpose languages (e.g., AnyLogic).

Interpreted languages must make substantial design compromises. Libraries require the user to learn a general programming language such as C or Java. AnyLogic appears to actually generate Java source code, much of which is specified visually, but it also requires the user to provide code in order to implement agent behaviors.

Eco utilizes a novel, compiled programming language which provides the potential for considerable performance optimization, while facilitating the use of computer-aided model design systems that reduce or eliminate the need for writing computer code.

Inventor

David Bruce Borenstein (http://www.davidbruceborenstein.com) is a 5^th year Ph.D. candidate in the Quantitative and Computational Biology program at Princeton University. He recently developed a novel programming language for spatial modeling, which has been used to explore community dynamics within microbial ecosystems. Before coming to Princeton, Mr. Borenstein worked as a research engineer at the Broad Institute of MIT and Harvard, where he created a natural-language data mining platform for genome sequences. In the 1990s, Mr. Borenstein set up an online quotation service, Quoteland.com, which he later sold to Awards International.

Intellectual Property and Technology Status

Patent protection is pending.

Industrial collaborators are sought for the further development and commercialization of this opportunity.

Contact

Laurie Tzodikov

Princeton University Office of Technology Licensing • (609) 258-7256• tzodikov@princeton.edu

Laurie Bagley

Princeton University Office of Technology Licensing • (609) 258-5579• lbagley@princeton.edu