Standard Information
- Vairables:
- CSold[k] new input introduced from C[k]*Sold[k]
- C[k] an input at time k (Sell Price)
- Sold[k] an input att time k (Amount Sold)
- xa[k] an input at time k (Purchasing from Supplier A)
- xb[k] an input at time k (Purchasing from Supplier B)
- xc[k] an input at time k (Purchasing from Supplier C)
- xd[k] an input at time k (Purchasing from Supplier D)
- y[k] the output at time k (Profit)
- Difference Equation (make-up parameter):
y[k] = y[k-1] + C[k]*Sold[k] - 500*xa[k-5] - 600*xb[k-3]
- 400*xc[k-7] - 550*xd[k-4]
- Modified Difference Equation (eliminate
nonlinearlity):
y[k] = y[k-1] + CSold[k] - 500*xa[k-5] - 600*xb[k-3]
- 400*xc[k-7] - 550*xd[k-4]
First Homework
- MATLAB file: hw1.m
- example input: graph
- simulated output: graph
- Ouput with Noise
- Noise Amplitude: 10% of the maximum
value of the output
- Noise Characteristic: uniform random noise
- simulated output + noise: graph
- ARX Method Result
- Ideal input:
y[k] = y[k-1] + CSold[k] - 500*xa[k-5] - 600*xb[k-3]
- 400*xc[k-7] - 550*xd[k-4]
- Ideal input + Noise:
y[k] = 0.7850*y[k-1] + 0.3307*CSold[k] + 1107*xa[k-5]
+ 3425*xb[k-3] + 969.4*xc[k-7] + 2558*xd[k-4]
- RARX Method Result
- Ideal input:
y[k] = y[k-1] + CSold[k] - 500*xa[k-5] - 600*xb[k-3]
- 400*xc[k-7] - 550*xd[k-4]
- Ideal input + Noise:
y[k] = 0.6200*y[k-1] + 8.8454*CSold[k] + 8197*xa[k-5]
+ 7381*xb[k-3] - 9119*xc[k-7] - 5871*xd[k-4]
- Comparision of the Quality of Parameters (2 Feb. 1996)
- MATLAB file: hw1_e.m
- Simulated output, graph, from
ideal input with
- the original parameter (yellow)
- the parameter from ARX method when there is noise (red)
- the parameter from RARX method when there is noise (green)