Seznam

Téma:Analogue Optimization Circuits
Vedoucí:Pavel Otta
Vypsáno jako:Práce v týmu a její organizace
Popis:In November 2014, Sergey Vichik and Francesco Borrelli from University of California in Berkeley published a paper in which solving of Linear Programming (LP) and Quadratic Programming (QP) by analogue circuits is presented. The main benefit of this approach is fast computation time reported to be in the few nanoseconds for small-sized problems as well as tiny energy consumption. LP or QP also arises in Model Predictive Control (MPC). This approach represents a new branch of our research currently aimed on fast algorithms for predictive control.

We are looking for enthusiastic theoretically or practically oriented students to develop the platform for conversion of optimization problem (LP and/or QP) into analogue circuit in Simscape/Matlab as well as circuit design, deployment and testing.
Pokyny:Following instructions are preliminary.

1) Review existing literature.
2) Make a platform for automatic design of electric circuit solving given optimization problems (LP or QP).
3) Test the designed electric circuits and compare results.

At least one software and one hardware oriented team member is required.
Literatura:Sergey Vichik, Francesco Borrelli, Solving linear and quadratic programs with an analog circuit. (http://www.sciencedirect.com/science/article/pii/S0098135414000131) and references therein.
Realizace:Simscape scheme, Matlab code, literature review notes
Vypsáno dne:10.02.2015
Max. počet studentů:4
Přihlášení studenti:
 

Upozornění: toto je závazné přihlášení. Zrušit ho může pouze vedoucí práce