Incentive-Compatible Mechanisms for Norm Monitoring in Open Multi-Agent Systems

Authors: Natasha Alechina, Joseph Y. Halpern, Ian A. Kash, Brian Logan

JAIR 2018 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show using simulations that our theoretical results, which apply to systems with a large number of agents, hold for multi-agent systems with as few as 1000 agents the system rapidly converges to the steady-state distribution of scrip tokens necessary to ensure monitoring and then remains close to the steady state. In this section, we use simulations to quantitatively evaluate the reasonableness of our theoretical results.
Researcher Affiliation Collaboration Natasha Alechina EMAIL University of Nottingham Nottingham, UK; Joseph Y. Halpern EMAIL Cornell University Ithaca, USA; Ian A. Kash EMAIL Microsoft Research Cambridge, UK; Brian Logan EMAIL University of Nottingham Nottingham, UK
Pseudocode No The paper describes steps in regular paragraph text without structured pseudocode or algorithm blocks. Mechanisms and strategies are explained descriptively.
Open Source Code No The paper does not mention providing access to source code for the methodology described.
Open Datasets No The paper describes simulated environments and token distributions but does not refer to or provide access information for any external or open datasets used for empirical evaluation.
Dataset Splits No The paper conducts simulations based on its theoretical model rather than using external datasets with explicit train/test/validation splits.
Hardware Specification No The paper describes simulations but does not specify any hardware details (e.g., GPU/CPU models, memory) used for running them.
Software Dependencies No The paper does not mention any specific software dependencies or their version numbers used in the simulations or for implementing the described mechanisms.
Experiment Setup Yes We report here results of simulations where there are twice as many tokens as agents (so the average number of tokens per agent shown to be a key parameter by KFH is two), agents used a threshold of k = 5, and, in the inadvertent setting, the probability of submtting a bad posting is b = 0.2. (These choices are arbitrary; similar results hold for other settings of the parameters.)