Is this project an undergraduate, graduate, or faculty project?
Undergraduate
individual
What campus are you from?
Daytona Beach
Authors' Class Standing
Julia Gorthey, Senior
Lead Presenter's Name
Julia Gorthey
Faculty Mentor Name
Byran Watson
Abstract
Swarm intelligence uses individual cooperation for improved community strength. However, some underwater swarm intelligence behaviors have capabilities for real-world applications that are underutilized. The application of biological-inspired design resource allocation to better methods of response management. Snapping shrimp colonies have a sophisticated approach to resource allocation under uncertainty. The snapping shrimp behavior will be the target for creating this algorithm for resource allocation based on a biological-inspired behavior. The research question this study will examine is how can the intruder response behavior of a snapping shrimp colony can be applied to model distributed resource allocation to show how factors such as local vs global knowledge and peer-to-peer communication can impact the threat elimination? The hypothesis for this work is by allocating only necessary resources and utilizing peer-to-peer communication, there will be a correlation to threat severity with agent count and sensing. The snapping shrimp behavior inspired an algorithm that was applied to an author-written Anylogic model simulating a multi-agent optimized wildfire response simulation. By applying the snapping shrimp resource allocation algorithm (SSRAA) to the agents in the Anylogic model, the results indicate there is an improved performance with less advanced agents with communication compared to more advanced agents with less communication. In the trials of the model, a number of agents with global knowledge and no agent communication successfully eliminated the threat but the agents with communication with the same sensing and agent number parameters are more or equally successful in the same tasks. The results prove that communication and agent coordination are more effective means to maintain limited resources. This work has a multitude of applications that can show how systems can better allocate their resources through agent communication with limited sensing through disturbed resource allocation.
Did this research project receive funding support from the Office of Undergraduate Research.
Yes, SURF
Biological Inspired Resource Allocation for Distributed Multi Agent System with limited knowledge
Swarm intelligence uses individual cooperation for improved community strength. However, some underwater swarm intelligence behaviors have capabilities for real-world applications that are underutilized. The application of biological-inspired design resource allocation to better methods of response management. Snapping shrimp colonies have a sophisticated approach to resource allocation under uncertainty. The snapping shrimp behavior will be the target for creating this algorithm for resource allocation based on a biological-inspired behavior. The research question this study will examine is how can the intruder response behavior of a snapping shrimp colony can be applied to model distributed resource allocation to show how factors such as local vs global knowledge and peer-to-peer communication can impact the threat elimination? The hypothesis for this work is by allocating only necessary resources and utilizing peer-to-peer communication, there will be a correlation to threat severity with agent count and sensing. The snapping shrimp behavior inspired an algorithm that was applied to an author-written Anylogic model simulating a multi-agent optimized wildfire response simulation. By applying the snapping shrimp resource allocation algorithm (SSRAA) to the agents in the Anylogic model, the results indicate there is an improved performance with less advanced agents with communication compared to more advanced agents with less communication. In the trials of the model, a number of agents with global knowledge and no agent communication successfully eliminated the threat but the agents with communication with the same sensing and agent number parameters are more or equally successful in the same tasks. The results prove that communication and agent coordination are more effective means to maintain limited resources. This work has a multitude of applications that can show how systems can better allocate their resources through agent communication with limited sensing through disturbed resource allocation.