Author Information

Grace Lynn GrattonFollow

Is this project an undergraduate, graduate, or faculty project?

Undergraduate

individual

What campus are you from?

Daytona Beach

Authors' Class Standing

Sophomore

Lead Presenter's Name

Grace Gratton

Faculty Mentor Name

Dr. Bryan Watson

Abstract

Multi-agent systems have several advantages over centralized systems of control and are becoming more heavily relied upon by systems. Consensus algorithms are a fundamental element of multi-agent systems, but current methods of consensus have a low robustness to faulted agents, endangering the integrity of the system. Biologically Inspired Design previously inspired a state-of-the-art consensus algorithm with resilience of up to 20% faulted agents within a population, called the Synchronous Hatching Consensus Algorithm. This article further tests the algorithm by implementing faulted agents reporting false negatives, rather than the original false positives, to see if the model remains robust. To test the proposed algorithm an Agent-Based, ANYLOGIC model was tested against 0, 1, 5, 10, 15, and 20 ‘false negative’ faulted agents across four varying environments. The robustness of the algorithm was measured by the time to consensus for 66% and the percentage of runs that failed to reach consensus. Three separate tests were conducted to determine if the model is robust to faulted agents reporting false positives. It was found that two of three tests deemed the model not robust. An uncertainty and sensitivity analysis was then performed to determine the uncertainty of failure to reach 66% consensus and the cause of uncertainty in the results. The total probability to not reach consensus was 59%, and the highest cause of this failure was the rate of change of the environment. The number of faulted agents had the second highest impact upon the uncertainty. Thus, we recommend this algorithm be implemented in an environment that quickly reaches its decision threshold and interacts with predominantly false positive faulted agents.

Did this research project receive funding support from the Office of Undergraduate Research.

Yes, SURF

Share

COinS
 

Evaluating a Biologically-Inspired Multi-Agent System Consensus Algorithm for Robustness to False Negatives

Multi-agent systems have several advantages over centralized systems of control and are becoming more heavily relied upon by systems. Consensus algorithms are a fundamental element of multi-agent systems, but current methods of consensus have a low robustness to faulted agents, endangering the integrity of the system. Biologically Inspired Design previously inspired a state-of-the-art consensus algorithm with resilience of up to 20% faulted agents within a population, called the Synchronous Hatching Consensus Algorithm. This article further tests the algorithm by implementing faulted agents reporting false negatives, rather than the original false positives, to see if the model remains robust. To test the proposed algorithm an Agent-Based, ANYLOGIC model was tested against 0, 1, 5, 10, 15, and 20 ‘false negative’ faulted agents across four varying environments. The robustness of the algorithm was measured by the time to consensus for 66% and the percentage of runs that failed to reach consensus. Three separate tests were conducted to determine if the model is robust to faulted agents reporting false positives. It was found that two of three tests deemed the model not robust. An uncertainty and sensitivity analysis was then performed to determine the uncertainty of failure to reach 66% consensus and the cause of uncertainty in the results. The total probability to not reach consensus was 59%, and the highest cause of this failure was the rate of change of the environment. The number of faulted agents had the second highest impact upon the uncertainty. Thus, we recommend this algorithm be implemented in an environment that quickly reaches its decision threshold and interacts with predominantly false positive faulted agents.

 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.