Use analytical solvers for 1-well potentials, numeric solver for multi-well potentials.

[pixelzoom]

We are currently using analytical or nearly-analytical solvers (for enery levels and wave functions) for the following potential types:

• Infinite Square
• Infinite Step
• Coulomb

Should we use analytical solvers for other potential types that are specific to the One Well screen? Those potentials are:

• Asymmetric Triangle
• Harmonic Oscillator
• Morse

And we would be using numerical solution (NumerovSolver) only for potentials that involve multiple wells:

• Finite Square
• Pöschl-Teller

Advantages:

• Highly responsive, no lag.
• Eases restrictions on parameter ranges.
• Can more easily demonstrate specific effects that are now difficult.
• Better encapsulation: code changes to one potential type do not affect others.

Disadvantages:

• More code.
• Cannot have analytical solution for multiple wells.
• Larger ranges causes wave function to extend beyond x-axis range.

[brettfiedler]

We voted in the meeting to proceed with analytic solutions for all potentials on the one well screen. We will switch to numerov for Double well and beyond.

For superposition screen, we should be able to use analytical solutions for all except for the double square.

[pixelzoom]

Here's a summary of the work that needs to be done for this issue.

These analytic solutions are complete:

• InfiniteSquareSolution
• InfiniteStepSolution
• CoulombSolution

These analytical solutions need to be created:

• AsymmetricTriangleSolution (including support for xOffset, yOffset)
• PosclTellerSolution (including support for numberOfWells, xOffset, yOffset, and electricField)

These analytical solutions need additional work to be "complete":

FiniteSquareSolution

• add computation of potentials to solve
• add support for numberOfWell, xOffset, yOffset, and electricField to createPotentialFunction

HarmonicOscillatorSolution

• add computation of potentials to solve
• add support for xOffset and yOffset to createPotentialFunction

MorseSolution

• add computation of potentials to solve
• add support for xOffset and yOffset to createPotentialFunction

[pixelzoom]

I'll need assistance from @veillette for the remaining work identified in #43 (comment).