The Solvent Selection Advisory System (SSAS) is a decision support system for aiding users in the preliminary selection of optimal solvents for carbon dioxide removal processes given different user specification and plant conditions. This paper describes an inference-network representation of the Solvent Selection Advisory System which has been previously implemented as a rule-based system. Two expert system shells, G2 and GDA from Gensym Corp. are used, and the implementations on the two shells are compared. Some advantages and disadvantages in the two representation approaches are discussed.