Large-Scale Decision Making Overview, published in Information Fusion

The research area of Large Group Decision Making (LGDM) has acquired growing importance in the last 5 to 6 six years, with many researchers around the globe developing models and solutions to support collective decisions at large scale.

A variant of this LGDM framework, called Large-Scale Decision Making is the subject of this article published in Information Fusion (Impact Factor: 10.716, ranked 3/134 by ISI-JCR in Computer Science: Artificial Intelligence), in which the DSRS Lab collaborated with scientists from University of Granada (Spain), Tianjin University (China) and Sichuan University (China).

Abstract:

The last decade witnessed tremendous developments in social media and e-democracy technologies. A funda- mental aspect in these paradigms is that the number of decision makers allowed to partake in a decision making event drastically increases. As a result Large Scale Decision Making (LSDM) has established itself as an emerging and rapidly developing research field, attracting comprehensive studies in the last decade. LSDM events are a complex class of decision making problems, in which multiple and highly diverse stakeholders are involved and the provided alternatives are assessed considering multiple criteria/attributes. Since some of the extant LSDM re- search was extended from group decision making scenarios, there is no established definition for a LSDM problem as of yet. We firstly propose a clear definition and characterization of LSDM events as a basis for characterizing this emerging family of decision frameworks. Secondly, a classification of LSDM literature is provided. Effectively solving an LSDM problem is usually a complex and challenging process, in which reaching a high consensus or accounting for the agreement or conflict relationships between participants becomes critical. Accordingly, we present a taxonomy and an overview of LSDM models, predicated on their key elements, i.e. the procedures and specific steps followed by the existing models: consensus measurement, subgroup clustering, behavior man- agement, and consensus building mechanisms. Finally, we provide a discussion in which we identify research challenges and propose future research directions under a triple perspective: key LSDM methodologies, AI and data fusion for LSDM, and innovative applications. The potential rise of AI-based LSDM is particularly highlighted in the discussion provided.

The article establishes a taxonomy around which existing LSDM literature is overviewed, based on the key elements/processes undertaken in LSDM models (e.g. consensus measurement and building, subgroup clustering, behavior management…) and the types of approaches adopted by different scholars for each of these key elements.

Fig. 7

 

Furthermore, the article provides a comprehensive discussion on the challenges and future research directions in LSDM, with special emphasis on (i) the potential role of AI and Data Fusion/Data Science technologies to improve these decision support models, and (ii) Innovative Areas of Application of LSDM.

 

Fig. 10

Article information:

Ru-Xi Ding, Iván Palomares, Xueqing Wang, Guo-Rui Yang, Bingsheng Liu, Yucheng Dong, Enrique Herrera-Viedma, Francisco Herrera. Large-Scale decision-making: Characterization, taxonomy, challenges and future directions from an Artificial Intelligence and applications perspective. Information Fusion. Volume 59, July 2020, Pages 84-102.

The following two tabs change content below.

i.palomares

Leave a Reply

Your email address will not be published. Required fields are marked *