Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
220 views

Optimal Control Homework 5

This document contains instructions for two homework problems on optimal control. Problem 1 involves finding the minimum time control of a linear system with state and input constraints. Problem 2 involves finding the minimum time control of the nonlinear Van der Pol system with varying input constraints. Students are asked to implement and solve the optimal control problems in Simulink, plotting results and analyzing the effects of changing initial conditions and constraints.

Uploaded by

jo Ten
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
220 views

Optimal Control Homework 5

This document contains instructions for two homework problems on optimal control. Problem 1 involves finding the minimum time control of a linear system with state and input constraints. Problem 2 involves finding the minimum time control of the nonlinear Van der Pol system with varying input constraints. Students are asked to implement and solve the optimal control problems in Simulink, plotting results and analyzing the effects of changing initial conditions and constraints.

Uploaded by

jo Ten
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

EE5321, Spring 2018

Homework Assignment 5: Min Time Constrained Input


Due 4/12/2018

Problem 1: Linear Bang-Bang Control (50 points)


𝑥̇ 1 = 𝑥2
Implement the system 𝑥̇ 2 = 𝑥3 in Simulink using a fixed solver with time step
𝑥̇ 3 = −𝑥1 − 𝑥2 − 𝑥3 + 𝑢
of 0.025 seconds. Impose a control (input) constraint of +/-1. Optimize the minimum time
𝑡
performance index 𝐽 = ∫0 𝑓 1 𝑑𝑡 with the final constraint of zero for all three states. (i.e. this
should result in a bang-bang type solution, but note that the switches might not be ‘crisp’). You
can use the options as options = optimset('Display', 'iter', 'TolCon', 1e-5,'TolFun', 1e-5,
'PlotFcns', 'optimplotx', 'MaxFunEvals', 4400.’Algorithm’,’interior-point’); and an initial control
of 0.01. Note you may have to adjust the maximum number of function evaluations to make the
final control look reasonable. You get to pick the initial guess at the optimal final time, but keep
it in the range of 1 to 6. Note you may have to try a few different final time guesses to ensure the
state constraints are satisfied at the final time.

a) Take the initial state to be (𝑥1 (0), 𝑥2 (0), 𝑥3 (0)) = (1,0,0) and find the optimal control
and optimal final time. Plot the final state and control time histories and show the optimal
final time. How many switches were needed?
b) Take the same initial states as a), but impose control constrained to +/- 0.75 and find the
optimal control and optimal final time. Plot the final state and control time histories and
show the optimal final time. How many switches were needed and did the switch times
change?
c) Take the same initial states as a), but impose control constrained to +/- 0.5 and find the
optimal control and optimal final time. Plot the final state and control time histories and
show the optimal final time. How many switches were needed? Did the approximate
switch times change and, if so, by how much?
d) Take the initial state to be (𝑥1 (0), 𝑥2 (0), 𝑥3 (0)) = (1,1,1) and find the optimal control
and optimal final time with the control constrained to +/- 1.0. Plot the final state and
control time histories and show the optimal final time. Did the number of switches
change?

Problem 2: Nonlinear Van der Pol (50 points)


For the Van der Pol system, implement the following in Simulink:

𝑥̇ 1 = 𝑥2 𝑥1 (0) = 0 𝑥1 (𝑡𝑓 ) = 3
𝑥̇ 2 = 𝜇(1 − 𝑥12 )𝑥2 − 𝑥1 + 𝑢 𝑥2 (0) = 0 𝑥2 (𝑡𝑓 ) = 0

𝑡
Minimize the performance index 𝐽 = ∫0 𝑓 1 𝑑𝑡 I suggest a fixed step integration of 0.02 seconds.
I suggest you have the initial control, u, to be 0.01 for all time 0:0.02:1. Also, I suggest you use
the following options:
options = optimset('Display','iter','TolCon',1e-4,'TolFun',1e-
4,'PlotFcns','optimplotx', 'MaxFunEvals',5000,'Algorithm','interior-
point');

a) Set the upper and lower bounds of the control to be +/-10, find the optimal final time and
plot the optimal state and control time histories when 𝜇 = 5.
b) Set the bounds to be +/-7 and rerun, plotting the optimal state and control time histories
and note the optimal final time.
c) Set the bounds to be +/-5 and rerun, plotting the optimal state and control time histories
and note the optimal final time.
d) Compare the optimal final times found for these three situations and note the trend. Make
comments about how the control time histories changed as well as the qualatative
differences between the three state time histories.

You might also like