Can in-silico approaches of human eating behaviours shed light on the key factors driving meal decisions? - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Can in-silico approaches of human eating behaviours shed light on the key factors driving meal decisions?

Résumé

Behind a seemingly simple task of selecting dishes to compose a meal in a collective setting, a complex decision-making process unfolds, influenced by a multitude of factors. We aimed to unravel the key determinants of meal composition choices within a university restaurant setting. However, conventional approaches alone are insufficient, we used an innovative methodology: Agent-Based Modeling. This novel approach combined the computational simulations with the assessment of guest's meal choices in natural consumption situations. The agent-based model incorporated various influential factors, including individual preferences for presented dishes, the social influence exerted by other agents within the simulated environment, and the impact of environmental cues, such as sensory perceptions and the waiting time for desired items. Simultaneously, real-world observations were conducted in two university restaurants during lunchtime, capturing the meal choices of 336 students. Anthropometric data and subjective decision criteria were collected through self- reported questionnaires, while the composition of each guest's table provided insights into interpersonal familiarity. By utilizing the collected characteristics of participants as initial parameters, simulations were conducted, comparing the choices of simulated agents with those of real participants. This comparative analysis allowed for the identification of specific values that defined the relative importance of preferences, social influences, and environmental factors in guests' decision-making processes. The results challenged initial perceptions, as the majority of respondents declared themselves immune to the influence of others. However, the models that yielded the least errors assigned considerable weight to the choices of others during the agents' decision-making process. Intriguingly, further exploration of the model revealed that not all dishes were equally susceptible to social influences. The variations in this parameter appeared to deferentially affect various components of the meal. This study highlights the heuristic value of Agent-Based Modeling in unraveling the intricate web of factors at play in meal composition decisions. By combining computational simulations with real-world observations, this methodological approach offers a promising perspective, shedding light on the underlying mechanisms of decision-making in natural consumption situations
Fichier non déposé

Dates et versions

hal-04542685 , version 1 (11-04-2024)

Licence

Paternité - Pas d'utilisation commerciale - Partage selon les Conditions Initiales

Identifiants

  • HAL Id : hal-04542685 , version 1

Citer

Damien Saveant, Mathieu Guimont, Patrick Taillandier, Nicolas Darcel. Can in-silico approaches of human eating behaviours shed light on the key factors driving meal decisions?. Benjamin Franklin & Lafayette Seminar, Jun 2023, Frejus (France), France. ⟨hal-04542685⟩
0 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More