May Revisions

[adowling2]

This issue tracks improvement tasks in May (after the practice run).

Slides:

• [ ] Add slide on why Pyomo
• [ ] Change slide template
• [ ] Add logos for other projects for more applications
• [ ] Add graphical overview/agenda

Plots:

• [x] Larger fonts
• [x] Consistent color scheme
• [x] Remove data from plots with only simulation results
• [x] Remove disturbance from plots

Modeling:

• [x] Add constant Cp assumption
• [x] Ub instead of Uc

Numpy:

• [ ] Add timing information (print out)

Pyomo tutorial:

• [x] Add objectives for each section to make the optimization problem really clear
• [x] Goals for the notebook
• [x] Emphasize square versus optimization problem
• [x] Show how to extract results from a Pyomo model

Parameter Estimation:

• [x] Remove vanilla Pyomo
• [x] Add excerpts of key code from tclab_pyomo.py file
• [x] Add LaTex for least squares objective

DoE exploratory analysis:

• [x] math and Pyomo code for sine wave parameterization
• [ ] Rerun on computer with k_aug installed.

New notebook on DoE extras:

• [ ] Can we do just half of the experiment?
• [ ] What if we combine the FIM from the two measurements? Does this fix the eigenvalue problem?
• [ ] What if we exclude CpS?
• [ ] What is we drop a fraction of the data, i.e., 1/2, 1/3, 1/4, 1/5, 1/6, 1/8, 1/9, 1/10?

DoE optimization:

• [x] Switch order, do A-optimality first
• [ ] What is the FIM of just the next experiment? Is the optimized experiment rank deficient?
• [x] How much do the A-optimality results depend on the values of theta?
• [x] Verify degrees of freedom for optimization problems.

General:

• [ ] Check links for all data files
• [ ] Run black for notebooks
• [x] Add exercise solutions to website

Pyomo.DoE improvements:

• [ ] Add restart if solver fails?
• [ ] Add warning about no measurement variables

Landing page:

• [x] Add names of presenter and contributors

[adowling2]

During the presentation, skip the numpy modeling. That will speed up the time.

[adowling2]

Pyomo.DoE exercise. D-opt:

Total number of variables............................:    44157
                     variables with only lower bounds:        4
                variables with lower and upper bounds:    24309
                     variables with only upper bounds:        0
Total number of equality constraints.................:    43250
Total number of inequality constraints...............:        0
        inequality constraints with only lower bounds:        0
   inequality constraints with lower and upper bounds:        0
        inequality constraints with only upper bounds:        0

907 degrees of freedom. Expected 900.

A-opt has 901 degrees of freedom.

Total number of variables............................:    44142
                     variables with only lower bounds:        1
                variables with lower and upper bounds:    24309
                     variables with only upper bounds:        0
Total number of equality constraints.................:    43241
Total number of inequality constraints...............:        0
        inequality constraints with only lower bounds:        0
   inequality constraints with lower and upper bounds:        0
        inequality constraints with only upper bounds:        0

[adowling2]

Looking further into the FIM:

fim : Size=16, Index=regression_parameters*regression_parameters
        Key                    : Lower : Value               : Upper : Fixed : Stale : Domain
                  ('Ua', 'Ua') :  None :  190391978.90971845 :  None : False : False :  Reals
                  ('Ua', 'Ub') :  None : -192132808.31251907 :  None : False : False :  Reals
             ('Ua', 'inv_CpH') :  None :  -8523487.709703343 :  None : False : False :  Reals
             ('Ua', 'inv_CpS') :  None : -1778.4067336132327 :  None : False : False :  Reals
                  ('Ub', 'Ua') :  None :                   0 :  None : False :  True :  Reals
                  ('Ub', 'Ub') :  None :  3634469242.1266756 :  None : False : False :  Reals
             ('Ub', 'inv_CpH') :  None :   61416421.37004702 :  None : False : False :  Reals
             ('Ub', 'inv_CpS') :  None :  23883.415279158395 :  None : False : False :  Reals
             ('inv_CpH', 'Ua') :  None :                   0 :  None : False :  True :  Reals
             ('inv_CpH', 'Ub') :  None :                   0 :  None : False :  True :  Reals
        ('inv_CpH', 'inv_CpH') :  None :   1586876.981660226 :  None : False : False :  Reals
        ('inv_CpH', 'inv_CpS') :  None :  19465.229008144302 :  None : False : False :  Reals
             ('inv_CpS', 'Ua') :  None :                   0 :  None : False :  True :  Reals
             ('inv_CpS', 'Ub') :  None :                   0 :  None : False :  True :  Reals
        ('inv_CpS', 'inv_CpH') :  None :                   0 :  None : False :  True :  Reals
        ('inv_CpS', 'inv_CpS') :  None :   4373.306916686016 :  None : False : False :  Reals

Fix lower diagonal elements to zero (because they are not calculated)?