-
Adapted one-sided differences gives a `NaN` *input* while receiving finite values from the function.
```julia
(jl_yIAkgW) pkg> st
Status `/tmp/jl_yIAkgW/Project.toml`
[26cc04aa] FiniteDifferen…
tpapp updated
2 months ago
-
The fact that `Finite n` represents a natural number in the *open* range [0, *n*), equivalent to the closed range [0, *n* − 1] rather than the closed range [0, *n*], seems to be a consistent source of…
-
For finite difference approximation of a derivative, a big O notation appears. The residual part (the error terms) is equal to `O(delta_x^n)` where `n` is the order of finite difference approximation.…
-
With increasing resolution the error is not showing a convergence order:
![AdvDiff_HP_convergence](https://github.com/underworldcode/underworld3/assets/55677727/f3c22513-b393-4939-9e02-bca3964b8ee0)
…
-
The TS is inconsistent wrt infinite ranges. It requires the end to be reachable in a finite number of increments, but it uses `unreachable` as a sentinel. It also has ye olde problem of difference typ…
-
We don't check that the boundary name used in the finite difference operator is actually in the field it is used with. This can cause problems.
-
### Powertrain Thermals
---
1. Run experiments to characterize PTN thermals (HV battery, LV battery, motor, inverter)
2. Build first principles models (finite difference may work if you can use lum…
-
Ceres docs advocate for using algorithmic differentiation instead of finite differences. Let's make one more IPy notebook example with the jacobian calculated using algopy, https://pythonhosted.org/al…
ev-br updated
9 years ago
-
I just cloned the master and set it up on my laptop for a clean start, and running the tests seems to do everything in working order, including giving the expected output
```
Test Summary: …
-
For the samplers which use derivatives (HMC, NUTS, ...), is it possible to use AD instead of finite differences? Even doing it manually (calculating the derivative with eg `ForwardDiff.jl` and supplyi…
tpapp updated
7 years ago