The increasing customer orientation of information technology applicat
ions has resulted in a stronger emphasis on providing query evaluation
services to the user community. Applications in electronic markets, d
ata warehousing and decision support systems arenas function in highly
dynamic environments serving users who demand increasing flexibility
in their interactions with the systems. These characteristics limit a
direct application of current query evaluation models. We propose the
development of a query evaluation subsystem (QUEM) that is equipped wi
th a knowledge base, a learning component and a decision support compo
nent to provide greater flexibility to the users and aid in managing s
ystem transitions. A prototype termed QUEST has been built and impleme
nted in a quasi-real world setting. The experimental results validate
the practical viability of the proposed architecture.