An official website of the United States government
Multiagent teams have been shown to be effective in many domains that require coordination among team members. However, finding valuable joint-actions becomes increasingly difficult in tightly-coupled domains where each agent’s performance depends on the actions of many other agents. Reward shaping partially addresses this challenge by deriving more “tuned" rewards to provide agents with additional feedback, but this approach still relies on agents ran- domly discovering suitable joint-actions. In this work, we introduce Counterfactual Agent Suggestions (CAS) as a method for injecting knowledge into an agent’s learning process within the confines of existing reward structures. We show that CAS enables agent teams to converge towards desired behaviors more reliably. We also show that improvement in team performance in the presence of suggestions extends to large teams and tightly-coupled domains.
more »
« less
On the Equality of Modal Damping Power and the Average Rate of Transient Energy Dissipation in a Multimachine Power System
- Award ID(s):
- 1739206
- PAR ID:
- 10356470
- Date Published:
- Journal Name:
- IEEE Control Systems Letters
- Volume:
- 6
- ISSN:
- 2475-1456
- Page Range / eLocation ID:
- 1531 to 1536
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This article adopts a relational perspective to demonstrate that characteristics of the dyadic relationship between supervisors and their employees are critical to understanding individual-level exploration—understood as the extent to which organizational members pursue new opportunities and experiment with changes to current practices. To this end, we introduce the concept of power framing—that is, whether the control over valued resources is emphasized as the ability to reward or to punish—and propose that supervisor power framing shapes employee exploration. In an experimental study, we demonstrate that reward (versus punishment) power framing increases employee exploration behavior and that this effect is mediated by perceived trustworthiness of the supervisor. In a second survey study, we replicate these findings in a field sample and show that the relationship between reward power framing and exploration depends on the degree to which the focal employee is sensitive to power characteristics (i.e., power distance orientation). This investigation advances scholarship on the microfoundations of exploration while also highlighting the ability of leaders to alter trustworthiness perceptions and induce employee exploration through power framing. Funding: This work was supported by a National Science Foundation CAREER Award from the Directorate for Social, Behavioral and Economic Sciences [Grant 1943688] granted to O. Schilke. Additional funding was provided by the Sauder School of Business, University of British Columbia. Supplemental Material: The online appendix is available at https://doi.org/10.1287/orsc.2023.1672 .more » « less
-
Energy-efficient IoT sensor nodes enable scalable monitoring of diverse physical environments, some of which are exposed to extreme and harsh operating conditions (such as heavy rain or strong movement). For reliable operation of such devices, certain variables must be adaptively adjusted. One of these variables is the transmission power of outgoing packets. In this work, we experimentally investigate how the movement of different bodies of water affects fluctuations in link quality and propose a model for predicting the received power. Once the received power is predicted, a transmitting node can adjust the transmission power to bring the received power to a desired level. Our model is based on minimum mean square estimation (MMSE) and leverages the received power statistics and the movement experienced by the nodes during communication. A disadvantage of MMSE is its dependence on matrix inversion, which is computationally intensive and difficult to implement on resource-constrained devices. We avoid this step and estimate the model parameters using gradient descent, which is much easier to implement. The model achieves an average prediction accuracy of 91% (SD = 1.7%) even with a small number of iterations.more » « less
-
null (Ed.)This paper studies seeded graph matching for power-law graphs. Assume that two edge-correlated graphs are independently edge-sampled from a common parent graph with a power-law degree distribution. A set of correctly matched vertex-pairs is chosen at random and revealed as initial seeds. Our goal is to use the seeds to recover the remaining latent vertex correspondence between the two graphs. Departing from the existing approaches that focus on the use of high-degree seeds in $$1$$-hop neighborhoods, we develop an efficient algorithm that exploits the low-degree seeds in suitably-defined D-hop neighborhoods. Specifically, we first match a set of vertex-pairs with appropriate degrees (which we refer to as the first slice) based on the number of low-degree seeds in their D-hop neighborhoods. This approach significantly reduces the number of initial seeds needed to trigger a cascading process to match the rest of graphs. Under the Chung-Lu random graph model with n vertices, max degree Θ(√n), and the power-law exponent 2<β<3, we show that as soon as D> 4-β/3-β, by optimally choosing the first slice, with high probability our algorithm can correctly match a constant fraction of the true pairs without any error, provided with only Ω((log n)4-β) initial seeds. Our result achieves an exponential reduction in the seed size requirement, as the best previously known result requires n1/2+ε seeds (for any small constant ε>0). Performance evaluation with synthetic and real data further corroborates the improved performance of our algorithm.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1109/LCSYS.2021.3121008
