Abstract
We recall that in linear programming small continuous changes of data can cause discontinuities, such as a sudden collapse or an explosion of the feasible set. This typically results in jumps of optimal solutions and the optimal value. The main objective of the report is to suggest a method for stable planning using linear and convex models. We show how to change data in the model so that a prescribed demand be satisfied and that, in the process of satisfying the demand, discontinuities do not occur.
2Research partly supported by Natural Sciences and Engineering Council of Canada and by le Ministère de l'Education du Québec (F.C.A.C.).
3Presented in part at the SIM-OP-IS '81, the Symposium on Operations Research held at Herceg Novi, Yugoslavia, September 30-October 2, 1981, and at the XI International Symposium on Mathematical Programming held in Bonn, Federal Republik of Germany, August 23-27, 1982.
2Research partly supported by Natural Sciences and Engineering Council of Canada and by le Ministère de l'Education du Québec (F.C.A.C.).
3Presented in part at the SIM-OP-IS '81, the Symposium on Operations Research held at Herceg Novi, Yugoslavia, September 30-October 2, 1981, and at the XI International Symposium on Mathematical Programming held in Bonn, Federal Republik of Germany, August 23-27, 1982.
Notes
2Research partly supported by Natural Sciences and Engineering Council of Canada and by le Ministère de l'Education du Québec (F.C.A.C.).
3Presented in part at the SIM-OP-IS '81, the Symposium on Operations Research held at Herceg Novi, Yugoslavia, September 30-October 2, 1981, and at the XI International Symposium on Mathematical Programming held in Bonn, Federal Republik of Germany, August 23-27, 1982.