Multi-SPLOT: Supporting Multi-user Conﬁgurations with Constraint Programming Sebastian Vel´ asquez-Guevara 1 , Gilberto Pedraza 1 , and Jaime Chavarriaga 2(B ) 1 Universidad Piloto de Colombia, Bogot´a, Colombia rlxsebas@gmail.com , gilberto-pedraza@unipiloto.edu.co 2 Universidad de los Andes, Bogot´a, Colombia ja.chavarriaga908@uniandes.edu.co Abstract. Nowadays, companies have moved from oﬀering a single product for all their clients, to oﬀer diﬀerent customized for each one. These companies provide Conﬁguration Systems where a user can decide and discard which features she wants in her ﬁnal product. However, although almost all of these systems support individual decisions, usu- ally they do not oﬀer an special support for decisions made by multiple users for the same product. This paper introduces Multi-SPLOT, a web- based Conﬁguration System that supports simultaneous decisions from multiple users. This system uses oﬀ-the-shelf solvers to determine if these decisions are not conﬂicting among them, and to propose solutions when the decisions of an user conﬂict with decisions of the others. This paper shows the design of the solution and details of its implementation using Angular, Firebase and the optimization library in Google App Script. Keywords: Feature models · Conﬁguration systems Multi-user conﬁguration 1 Introduction Installing and conﬁguring complex applications may take many months and involve multiple users. For instance, deploying and conﬁguring a system such as SAP takes up an average between 6 to 12 months [8]. Diverse users, with diﬀerent areas of knowledge and requirements (e.g. engineers, accountants and manufac- turing experts) must coordinate their work to achieve a successfull installation. In a Conﬁguration Systems (a.k.a. a Conﬁgurator), users select options that they want to include and enter values for the diverse parameters required by the product. The Software checks if the user selections do not have conﬂicts among them, i.e. that they can be included in the product at the same time, and oﬀers recommendations to the user if it ﬁnds some problems [7]. However, these systems usually consider all the decisions as if they were made by the c  Springer Nature Switzerland AG 2018 H. Florez et al. (Eds.): ICAI 2018, CCIS 942, pp. 364–378, 2018. https://doi.org/10.1007/978-3-030-01535-0_27