SPEED OF CONVERGENCE AND STOPPING RULES IN AN ITERATIVE PLANNING PROCEDURE FOR NONCONVEX ECONOMIES

A new planning procedure is proposed for economies with free disposal in production and monotonic social welfare functions. No convexity of any kind is required. The procedure takes the novel approach of searching for an optimum in the region of the unknown production set that the central planning board knows the least about, instead of near the most preferred part of the current estimate of the production set. This makes it possible to calculate an upper bound on the speed of convergence, estimate the distance between the current tentative plan and the optimal plans, and define sensible stopping rules. © 1993 Academic Press, Inc.