The growth of online marketplaces as multi-sided digital platforms has created a decision-making environment for sellers that differs fundamentally from traditional retail, combining a richer behavioral data trace with a narrower set of controllable parameters. This study addresses the resulting gap between the volume of available behavioral data and the absence of an integrated decision support architecture that connects this data to the several distinct types of decisions a seller must make. The aim of the study is to develop an enterprise-architecture model, in the ArchiMate notation, of a decision support system (DSS) for marketplace sellers, together with the mathematical models underlying its analytical and decision-support modules. The study applies methods of systematic literature analysis, enterprise architecture modeling, and mathematical modeling, including discrete-time hazard models and the Kolmogorov-Gabor polynomial. The resulting architecture comprises Businessand Application-layer ArchiMate models spanning four seller roles and four decision scenarios, supported by a shared analytical pipeline and explicit integration with the marketplace platform and external market-data providers. The proposed architecture and mathematical apparatus can be used as a reference design for information systems supporting marketplace sellers and as a basis for further empirical validation.