I'll use the Java API here, but the same approach works in at least some of the other APIs (including C++ and presumably C). The IloCplex class has methods named getBasisStatus, getBasisStatuses, setBasisStatuses (no singular version of this one) and setStart. Assuming that you are using a simplex algorithm (or barrier with crossover), one approach is to use getBasisStatuses to get and store the final basis of the solved cut generating LP, map them to the corresponding variables and constraints of the next LP and use setBasisStatuses. Another possibility is to get the primal and dual solution of the first LP, along with reduced costs, using the normal getters, map them to the new LP and use setStart to provide a (partial?) starting solution.
I believe the lpex6 example supplied with CPLEX Studio illustrates the use of setBasisStatuses.
------------------------------
Paul Rubin
Professor Emeritus
Michigan State University
------------------------------