59 - Implement uncertainty measure computation

[timeverettadams]

Resolves #59

Proposed Changes

• Implement uncertainty measure frequency response computation for various LFT uncertainty models (additive, multiplicative input/output, inverse additive, inverse multiplicative input/output).
  • Conversion of LFT uncertainty model frequency response to equivalent additive uncertainty model response.
  • Computation of uncertainty model representation with minimal size of Delta block.
  • Direct computation of uncertainty measure for additive uncertainty model.
  • Plotting uncertainty measure frequency response.
• Vectorize computation of uncertainty residuals across frequency using scipy array manipulation to speed up and simplify residual computation.
• Add computation of optimal uncertainty weight responses using full matrices (as opposed to just scalar and diagonal structured matrices).
• Replace the uncertainty weight magnitude plot with a decoupled singular value plot.
  • The previous iteration plots the magnitude, which is not meaningful in general (e.g., for full uncertainty weights), and it couples the plotting of the left and right uncertainty weights together. The new singular value plot allows for separate plotting of the weights while showing a more meaningful quantity for the size of the uncertainty set.
• Restructure the uncertainty weight fit code.

Checklist

• Write unit tests
• Write examples in docstrings
• Update Sphinx documentation
• Bump version number and date in pyproject.toml, doc/conf.py,
  README.rst, and CITATION.cff.