Scholarlink

<p>The Real-Time Geo-Spatial Consensus System is a spatial decision support system designed to facilitate the administration of spatial questionnaires to a panel of experts, to facilitate spatial consensus on territorial contexts. The platform enables experts to respond anonymously to one or more questions by placing spatial points, submitting comments, and reviewing results in real-time, thereby fostering active collaboration throughout the process. However, as documented in the scientific literature, experts often face competing commitments, which can result in inconsistent participation in sessions and limited collaboration with others. This paper addresses this challenge by incorporating a “super expert” within the platform, represented by a generative pre-trained transformer model. This model is integrated into the platform with a computational algorithm to perform multiple tasks, generating responses by referencing and analyzing the contributions of other participants. Findings from a transportation case study reveal that incorporating the model improves the efficiency of expert collaboration by shortening the time needed to reach a consensus of 63% while also providing relevant information to support the decision-making process.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrating generative pre-trained transformers in spatial decision support systems to facilitate expert consensus

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月31日

查看全文
  • Yuri Calleo,
  • Francesco Pilla

摘要

The Real-Time Geo-Spatial Consensus System is a spatial decision support system designed to facilitate the administration of spatial questionnaires to a panel of experts, to facilitate spatial consensus on territorial contexts. The platform enables experts to respond anonymously to one or more questions by placing spatial points, submitting comments, and reviewing results in real-time, thereby fostering active collaboration throughout the process. However, as documented in the scientific literature, experts often face competing commitments, which can result in inconsistent participation in sessions and limited collaboration with others. This paper addresses this challenge by incorporating a “super expert” within the platform, represented by a generative pre-trained transformer model. This model is integrated into the platform with a computational algorithm to perform multiple tasks, generating responses by referencing and analyzing the contributions of other participants. Findings from a transportation case study reveal that incorporating the model improves the efficiency of expert collaboration by shortening the time needed to reach a consensus of 63% while also providing relevant information to support the decision-making process.