Matthew Bohr, Ambrose A. Adegbege
This work focuses on creating an analog circuit using complementary metal-oxide semiconductor (CMOS) technology to aid in solving quadratic programming problems (QPP). QPP Circuits using op-amp and operational transconductance amplifiers (OTA) were also investigated. Quadratic programming is type of convex optimization found in the field of control and it entails minimizing a quadratic function subject to constraints. The circuit will ultimately be for control related applications due to the presence of QPPs in this field.
The aim of the research is to design multiple circuits capable of performing some mathematical operation such as addition, integration, or vector-matrix multiplication. Particular attention is paid to using CMOS transistors to re- alize these circuits given their potential for a VLSI implementation. Specific nodes within the circuit converge towards an equilibrium voltage, in turn pro- viding the values of the optimal vector of optimization variables that minimi- zes the quadratic program.