A tutorial on C/GMRES and automatic code generation for nonlinear model predictive control

This paper provides a tutorial on the continuation/ GMRES method (C/GMRES), which is a real-time optimization algorithm tailored for nonlinear model predictive control (NMPC) and a Maple version of AutoGenU, which is an automatic code generation system utilizing a symbolic computation language to simulate NMPC with C/GMRES. Once such settings of NMPC as the state equation and performance index are specified in a Maple worksheet, a problem-dependent C source file is automatically generated by symbolic computation, compiled, and executed. The worksheet also loads data files of simulation results and plots their time histories. This paper describes the usage and settings of AutoGenU for Maple.