Selecting the best taste: a group decision-making application to chocolates design N´ uria Agell and Germ´ an S´ anchez ESADE Business School Universitat Ramon Llull Av. Pedralbes 62, Barcelona, Spain Email: nuria.agell@esade.edu M´ onica S´ anchez and Francisco Javier Ruiz Universitat Politcnica de Catalunya, UPC-Barcelona Tech. Jordi Girona, 1-3, Barcelona, Spain Email: {monica.sanchez, francisco.ruiz}@upc.edu Abstract—Creation and design of products based on human sensory perceptions, such as color, smell or taste, require the participation of professionals or experts with highly developed sensory abilities. When a group of experts is involved in such creative process as a team, consensus and group decision-making (GDM) techniques able to deal with qualitative descriptions and uncertainty, can be required. In this paper we consider a methodology based upon qualitative reasoning techniques for representing and synthesizing the information given by a group of experts in order to capture the sensorial aspects of the alternatives. A real application of the proposed GDM method to chocolates design has been conducted throughout 2012 with the Chocolate Chef Oriol Balaguer’s team. We present the results obtained by applying the proposed methodology to aggregate experts’ opinion during a creative session. In this session, some members of Oriol Balaguer’s team tested and evaluated combi- nations of black chocolate with six different fruits, considered as alternatives, to select the best combination for the design and creation of a new cake. I. I NTRODUCTION Creative industries with specialized professionals that use their sensory abilities for creating and designing their products, generate enormous challenges in managing and replicating these skills [13], [16]. The food, beverage, perfume, and other creative industries continuously face dilemmas in modeling the cognitive ability and processes of these highly specialized individuals [3]. The characteristics of these kinds of processes are not purely functional and physical, but arise from highly subjective perceptive and cognitive aspects. In fact, these processes entail special abilities in human senses such as taste (beverages, food, chocolate, etc.), smell (perfume, toiletries, etc.), vision (color and shape of products), touch (fabrics, materials finishes, etc.), or hearing (e.g. acoustic diagnosis) [9], [4], [16]. When a group of experts is involved in such a creative process as a team, consensus and group decision- making techniques able to deal with qualitative descriptions and uncertainty are needed. Group decision-making (GDM) systems have been broadly investigated and used in different application areas, such as engineering, economics and management sciences [12], [14]. Their main interest is to allow decision makers (DMs) to reach a consensual solution without the need of neither their interaction nor a moderator, which leads to a less pressured and more anonymous participation [2], [11]. In general a team of experts has to rank a set of alternatives that are characterized by a set of conflicting attributes [10], [11], [17]. However, in the study carried out in this paper, and due to the sensorial condition of the alternatives to be ranked, the DMs evaluate each alternative as a whole, by means of qualitative linguistic labels corresponding to ordinal values [6], [7], [8]. Different levels of precision are considered to draw the distinctions required by the experts’ sensorial decision processes. In this paper we present an adaptation of the GDM method presented in [1] to capture the sensorial aspects of the alternatives. The method presented in this paper is based upon qualitative reasoning techniques for representing and synthesizing the information given by a group of experts and it can work at different precision levels. The scientific goals of the proposed methodology are representing, learning and predicting processes in a domain where perceived features are inherently vague, qualitative, imprecise, and often metaphor- ical. We present a real application that has been conducted throughout 2012 with Oriol Balaguer, a chocolate chef ac- tively involved in creative cuisine concept. Frequently, highly recognized chocolate chef Oriol Balaguer before deciding the launch of a new product explores the sensorial perception of the newly created product among the members of his team. The paper is structured as follows. First, the considered methodology of GDM is given in Section 2, where the absolute order-of-magnitude linguistic labels model is introduced and a distance among linguistic labels is considered. Then, in Section 3, the description of the problem of chocolates design is presented and the obtained experimental results are given. Finally, Section 4 contains the main conclusions and lines of future research. II. METHODOLOGY The methodology used in this paper is an adaptation of the method presented in [1] to capture the sensorial aspects of the alternatives to be ranked. For this reason, we consider an order-of magnitude qualitative model with 5 basic linguistic labels totally ordered as a chain: B 1 < ··· <B 5 to describe the alternatives A 1 , ··· ,A n . 939 978-1-4799-0348-1/13/$31.00 ©2013 IEEE