Indexing Structures play a critical role in representing mathematical general structures. In spite of the importance of indexing structure, there is few research about this field. We suggest new index structures in the UNIK-OPT environment. UNIK-OPT is a knowledge-assisted optimization model formulator. Because this system use domain knowledge as well as modeling knowledge, we could consutruct effective and efficient index structure using this characteristics. The suggested index structure is base index, secondary index and derived index. To test and confirm the capability of these structure, we use 25 examples which has been used in the field of modeling languages. And we represent the traveling salesman problem using these index structures.
In view of model maintenance, we developed index change propagation process in response to change in the problem domain being modelled.