Optimization-based design of engineering systems.
Allows multiple objectives and soft and hard constraints. Handles functional (semi-infinite) type specifications. Interactive tradeoff exploration. Easily coupled to user-supplied simulators; interface with MATLAB is available from the authors.
The given problem is transformed into a constrained, weighted minimax problem, with a different transformation depending on whether the soft/hard constraints are satisfied. The minimax problem is solved by using feasible sequential quadratic programming (see the entry on FSQP/CFSQP), extended to accommodate semi-infinite constraints. The user interactively tightens or relaxes specifications by adjusting "good values" and "bad values" (good/bad "curves" in the case of functional specifications), which determine the weights in the minimax problems.
The software runs only on Sun computers. It makes use of dynamic loading. A detailed manual is available. The package is available free of charge to academic institutions with FTP access. An elaborate graphical user interface with graphical input capability is under development.
Contact:
Prof. André L. Tits Ms. Carolyn Garrett Electrical Engineering Dept. and Office of Technology Liaison Institute for Systems Research 4312 Knox Road University of Maryland University of Maryland College Park, MD 20742 College Park, MD 20742 U.S.A. U.S.A. Phone: (301) 405-3669 Phone: (301) 405-4209 Fax: (301) 405-6707 Fax: (301) 314-9871 andre@src.umd.edu cg54@umail.umd.edu
M. K. H. Fan, A. L. Tits, J. Zhou, L.-S. Wang, and J. Koninckx, CONSOLE User's Manual, Version 1.1, TR 87-212r1, Systems Research Center, 1990.
M. K. H. Fan, L.-S. Wang, J. Koninckx and A. L. Tits, Software package for optimization-based design with user-supplied simulators, IEEE Control Systems Magazine 9 (1989), pp. 66--71.
[ Optimization Software Guide | OTC Home Page | NEOS Server | NEOS Guide ]