[WIP] pseudocode for VectorPeps and exact time evolution given gradient

[maddiecain]

initial work on VectorPeps for time-evolution benchmarking. Much is pseudocode because I haven't had time yet to figure out the correct syntax

[GiggleLiu]

It looks like you are mostly on the right track, except the test.

You use two methods to compute the time evolution, one is using the API in Yao, another is matrix exponentiation. I think the second method should be repeatedly doing the following calculation

• Compute x := dθ/dt,
• Do a single step time evolution θ = θ + x*Δt.