The best-of-N problem in collective decision making is complex especially when the number of available alternatives is larger than a few, and no alternative distinctly shines over the others. Additionally, if the quality of the available alternatives is not a priori known and noisy, errors in the quality estimation may lead to the premature selection of sub-optimal alternatives. A typical speed-accuracy trade-off must be faced, which is hardened by the presence of several alternatives to be analyzed in parallel. In this study, we transform a one-shot best-of-N decision problem in a sequence of simpler decisions between a small number of alternatives, by organizing the decision problem in a hierarchy of choices. To this end, we construct an m-ary tree where the leaves represent the available alternatives, and high-level nodes group the low-level ones to present a low-dimension decision problem. Results from multi-agent simulations in both a fully-connected topology and in a spatial decision problem demonstrate that the sequential collective decisions can be parameterized to maximize speed and accuracy against different decision problems. A further improvement relies on an adaptive approach that automatically tunes the system parameters.

Best-of-N Collective Decisions on a Hierarchy / Oddi, F.; Cristofaro, A.; Trianni, V.. - (2022), pp. 66-78. - LECTURE NOTES IN COMPUTER SCIENCE. [10.1007/978-3-031-20176-9_6].

Best-of-N Collective Decisions on a Hierarchy

Oddi F.
;
Cristofaro A.;
2022

Abstract

The best-of-N problem in collective decision making is complex especially when the number of available alternatives is larger than a few, and no alternative distinctly shines over the others. Additionally, if the quality of the available alternatives is not a priori known and noisy, errors in the quality estimation may lead to the premature selection of sub-optimal alternatives. A typical speed-accuracy trade-off must be faced, which is hardened by the presence of several alternatives to be analyzed in parallel. In this study, we transform a one-shot best-of-N decision problem in a sequence of simpler decisions between a small number of alternatives, by organizing the decision problem in a hierarchy of choices. To this end, we construct an m-ary tree where the leaves represent the available alternatives, and high-level nodes group the low-level ones to present a low-dimension decision problem. Results from multi-agent simulations in both a fully-connected topology and in a spatial decision problem demonstrate that the sequential collective decisions can be parameterized to maximize speed and accuracy against different decision problems. A further improvement relies on an adaptive approach that automatically tunes the system parameters.
2022
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
978-3-031-20175-2
978-3-031-20176-9
best-of-N problem; collective decision making; hierarchy
02 Pubblicazione su volume::02a Capitolo o Articolo
Best-of-N Collective Decisions on a Hierarchy / Oddi, F.; Cristofaro, A.; Trianni, V.. - (2022), pp. 66-78. - LECTURE NOTES IN COMPUTER SCIENCE. [10.1007/978-3-031-20176-9_6].
File allegati a questo prodotto
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/1664347
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact