Search Strategies in Evolutionary Multi-Agent Systems: The Effect of Cooperation and Reward on Solution Quality

2011 ◽  
Vol 133 (6) ◽  
Author(s):  
Lindsay Hanna Landry ◽  
Jonathan Cagan
Keyword(s):  
Fitness Function ◽  
Three Dimensional ◽  
Complex Problem ◽  
Multi Agent Systems ◽  
Solution Quality ◽  
Agent Systems ◽  
Fitness Functions ◽  
Engineering Design Problems ◽  
Multi Agent ◽  
Improved Performance

Cooperation and reward of strategic agents in an evolutionary optimization framework is explored in order to better solve engineering design problems. Agents in this Evolutionary Multi-Agent Systems (EMAS) framework rely on one another to better their performance, but also vie for the opportunity to reproduce. The level of cooperation and reward is varied by altering the amount of interaction between agents and the fitness function describing their evolution. The effect of each variable is measured using the problem objective function as a metric. Increasing the amount of cooperation in the evolving team is shown to lead to improved performance for several multimodal and complex numerical optimization and three-dimensional layout problems. However, fitness functions that utilize team-based rewards are found to be inferior to those that reward on an individual basis. The performance trends for different fitness functions and levels of cooperation remain when EMAS is applied to the more complex problem of three-dimensional packing as well.

The Effect of Reward Interdependence of Strategies in Evolutionary Multi-Agent Systems

2009 ◽  
Author(s):  
Lindsay Hanna ◽  
Jonathan Cagan
Keyword(s):  
Organizational Behavior ◽  
System Analysis ◽  
Three Dimensional ◽  
Computational Design ◽  
Design Teams ◽  
Multi Agent Systems ◽  
Agent Systems ◽  
Effective Strategies ◽  
Multi Agent ◽  
Potential Impact

This paper explores the effect of reward interdependence of strategies in a cooperative evolving team on the performance of the team. Experiments extending the Evolutionary Multi-Agent Systems (EMAS) framework to three dimensional layout are designed which examine the effect of rewarding helpful, in addition to effective strategies on the convergence of the system. Analysis of communication within the system suggests that some agents (strategies) are more effective at creating helpful solutions than creating good solutions. Despite their potential impact as enablers for other strategies, when their efforts were not rewarded, these assistant agent types were quickly removed from the population. When reward was interdependent, however, this secondary group of helpful agents remained in the population longer. As a result, effective communication channels remained open and the system converged more quickly. The results support conclusions of organizational behavior experimentation and computational modeling. The implications of this study are twofold. First, computational design teams may be made more effective by recognizing and rewarding indirect contributions of some strategies to the success of others. Secondly, EMAS may provide a platform for predicting the effectiveness of different reward structures given a set of strategies in both human and computational teams.

scholarly journals SYSTEM ANALYSIS IN HIERARCHICAL INTELLIGENT MULTI-AGENT SYSTEMS

2021 ◽  
pp. 33-46
Author(s):  
V. S. Simankov ◽  
Yu. V. Dubenko
Keyword(s):  
Reinforcement Learning ◽  
System Analysis ◽  
Three Dimensional ◽  
Label Propagation ◽  
Multi Agent System ◽  
Multi Agent Systems ◽  
Agent Systems ◽  
Hierarchical Reinforcement Learning ◽  
Multi Agent ◽  
Weight Coefficients

The system analysis of the hierarchical intelligent multi-agent system in general, as well as its main structural unit, the intelligent agent, its major subsystems identified. As part of the analysis of the computer vision subsystem, it was concluded that the considered sources have insufficiently worked out issues related to the processing of occlusions, with the automation of the process of reconstruction of three-dimensional scenes, with the implementation of the processing of an unstructured set of images. The structure of the block for the reconstruction of three-dimensional scenes is proposed, the implementation of which is aimed at eliminating the indicated problems characteristic of the machine vision subsystem. The analysis of the main methods of implementing unsupervised learning is carried out, based on the results of which it is concluded that it is advisable to use reinforcement learning when implementing systems of this type. Such types of reinforcement learning as hierarchical reinforcement learning and multi-agent reinforcement learning are considered. A method for segmentation of macro actions is proposed, based on the implementation of clustering by the method of label propagation, in which the number of transitions is formalized in the form of weight coefficients of edges.

scholarly journals A Fast Non-Linear Symmetry Approach for Guaranteed Consensus in Network of Multi-Agent Systems

Symmetry ◽  
2020 ◽  
Vol 12 (10) ◽  
pp. 1692
Author(s):  
Rawad Abdulghafor ◽  
Sultan Almotairi
Keyword(s):  
Smart Grids ◽  
Autonomous Agents ◽  
Linear Models ◽  
Low Complexity ◽  
Complex Problem ◽  
Multi Agent Systems ◽  
Quadratic Operator ◽  
Agent Systems ◽  
Non Linear ◽  
Multi Agent

There has been tremendous work on multi-agent systems (MAS) in recent years. MAS consist of multiple autonomous agents that interact with each order to solve a complex problem. Several applications of MAS can be found in computer networks, smart grids, and the modeling of complex systems. Despite numerous benefits, a significant challenge for MAS is achieving a consensus among agents in a shared target task, which is difficult without applying certain mathematical equations. Non-linear models offer better possibility of resolving consensus for MAS; however, existing non-linear models are considerably complicated and present no guarantees for achieving consensus. This paper proposes a non-linear mathematical model of semi symmetry quadratic operator (SSQO) in order to resolve the issue of consensus in networks of MAS. The model is based on stochastic quadratic operator theory, with added new notations. An important feature for the proposed model is low complexity, fast consensus, and a guaranteed capability to reach a consensus. We present an evaluation of the proposed SSQO model and comparison with other existing models. We demonstrate that an average consensus can be achieved with our model in addition to the emulation effects for the MAS consensus.

Multi-Agent Systems in Three-Dimensional Protein Structure Prediction

2017 ◽  
pp. 241-278 ◽  
Author(s):  
Leonardo de Lima Corrêa ◽  
Márcio Dorn
Keyword(s):  
Protein Structure ◽  
Protein Structure Prediction ◽  
Structure Prediction ◽  
Protein Structures ◽  
Three Dimensional ◽  
Multi Agent Systems ◽  
Agent Based ◽  
Agent Systems ◽  
Tertiary Protein Structure ◽  
Multi Agent

Tertiary protein structure prediction in silico is currently a challenging problem in Structural Bioinformatics and can be classified according to the computational complexity theory as an NP-hard problem. Determining the 3-D structure of a protein is both experimentally expensive, and time-consuming. The agent-based paradigm has been shown a useful technique for the applications that have repetitive and time-consuming activities, knowledge share and management, such as integration of different knowledge sources and modeling of complex systems, supporting a great variety of domains. This chapter provides an integrated view and insights about the protein structure prediction area concerned to the usage, application and implementation of multi-agent systems to predict the protein structures or to support and coordinate the existing predictors, as well as it is advantages, issues, needs, and demands. It is noteworthy that there is a great need for works related to multi-agent and agent-based paradigms applied to the problem due to their excellent suitability to the problem.

scholarly journals Generative Attention Networks for Multi-Agent Behavioral Modeling

2020 ◽  
Vol 34 (05) ◽  
pp. 7195-7202
Author(s):  
Guangyu Li ◽  
Bo Jiang ◽  
Hao Zhu ◽  
Zhengping Che ◽  
Yan Liu
Keyword(s):  
Interaction Analysis ◽  
Heterogeneous Systems ◽  
Generative Models ◽  
Behavior Prediction ◽  
Multi Agent Systems ◽  
Agent Systems ◽  
Multi Agent ◽  
Interaction Types ◽  
Improved Performance ◽  
Modeling Behavior

Understanding and modeling behavior of multi-agent systems is a central step for artificial intelligence. Here we present a deep generative model which captures behavior generating process of multi-agent systems, supports accurate predictions and inference, infers how agents interact in a complex system, as well as identifies agent groups and interaction types. Built upon advances in deep generative models and a novel attention mechanism, our model can learn interactions in highly heterogeneous systems with linear complexity in the number of agents. We apply this model to three multi-agent systems in different domains and evaluate performance on a diverse set of tasks including behavior prediction, interaction analysis and system identification. Experimental results demonstrate its ability to model multi-agent systems, yielding improved performance over competitive baselines. We also show the model can successfully identify agent groups and interaction types in these systems. Our model offers new opportunities to predict complex multi-agent behaviors and takes a step forward in understanding interactions in multi-agent systems.

Three-Dimensional Dynamic Formation Control of Multi-Agent Systems Using Rigid Graphs

2015 ◽  
Vol 137 (11) ◽  
Author(s):  
Pengpeng Zhang ◽  
Marcio de Queiroz ◽  
Xiaoyu Cai
Keyword(s):  
Formation Control ◽  
Three Dimensional ◽  
Multi Agent Systems ◽  
Agent Systems ◽  
3D Space ◽  
Multi Agent ◽  
Error Dynamics ◽  
Integrator Model ◽  
Double Integrator ◽  
Dynamic Formation

In this paper, we consider the problem of formation control of multi-agent systems in three-dimensional (3D) space, where the desired formation is dynamic. This is motivated by applications where the formation size and/or geometric shape needs to vary in time. Using a single-integrator model and rigid graph theory, we propose a new control law that exponentially stabilizes the origin of the nonlinear, interagent distance error dynamics and ensures tracking of the desired, 3D time-varying formation. Extensions to the formation maneuvering problem and double-integrator model are also discussed. The formation control is illustrated with a simulation of eight agents forming a dynamic cube.

Sign in / Sign up

Export Citation Format

Share Document