-
-
Notifications
You must be signed in to change notification settings - Fork 18
/
milne.cljc
69 lines (58 loc) · 3.01 KB
/
milne.cljc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
#_"SPDX-License-Identifier: GPL-3.0"
(ns emmy.numerical.quadrature.milne
(:require [emmy.numerical.quadrature.common :as qc]
[emmy.numerical.quadrature.midpoint :as qm]
[emmy.polynomial.richardson :as pr]
[emmy.util.stream :as us]))
;; ## Milne's Rule
;;
;; This numerical integration method is an [open Newton-Cotes
;; formula](https://en.wikipedia.org/wiki/Newton%E2%80%93Cotes_formulas#Open_Newton%E2%80%93Cotes_formulas);
;; for each integral slice, Milne's rule samples three interior points (not the
;; endpoints!) and combines them into an area estimate for this slice using the
;; following formula:
;;
;; $${{4h} \over 3} (2f_1 - f_2 + 2f_3)$$
;;
;; Given a window of $(a, b)$ and a "step size" of $h = {{b - a} \over 3}$. The
;; point $f_i$ is the point $i$ steps into the window.
;;
;; There is a simpler way to understand this! Milne's method is, in fact, just
;; the midpoint method (see `midpoint.cljc`), subject to a single refinement
;; of "Richardson extrapolation".
;;
;; The test namespace contains a symbolic proof that the Richardson-extrapolated
;; Midpoint method is equivalent to using the formula above to calculate
;; Milne's rule directly.
(defn milne-sequence
"Returns a (lazy) sequence of successively refined estimates of the integral of
`f` over the open interval $(a, b)$ using [Milne's
rule](https://en.wikipedia.org/wiki/Newton%E2%80%93Cotes_formulas#Open_Newton%E2%80%93Cotes_formulas).
Milne's rule is equivalent to the midpoint method subject to one refinement of
Richardson extrapolation.
Returns estimates with $n, 2n, 4n, ...$ slices, geometrically increasing by a
factor of 2 with each estimate.
## Optional arguments:
If supplied, `:n` (default 1) specifies the initial number of slices to use.
NOTE: the Midpoint method is able to reuse function evaluations as its windows
narrow _only_ when increasing the number of integration slices by 3. Milne's
method increases the number of slices geometrically by a factor of 2 each
time, so it will never hit the incremental path. You may want to memoize your
function before calling [[milne-sequence]]."
([f a b] (milne-sequence f a b {:n 1}))
([f a b {:keys [n] :or {n 1} :as opts}]
{:pre [(number? n)]}
(-> (qm/midpoint-sequence f a b (assoc opts :n (us/powers 2 n)))
(pr/richardson-column 1 2 2 2))))
(qc/defintegrator integral
"Returns an estimate of the integral of `f` over the open interval $(a, b)$
using [Milne's
rule](https://en.wikipedia.org/wiki/Newton%E2%80%93Cotes_formulas#Open_Newton%E2%80%93Cotes_formulas)
with $1, 2, 4 ... 2^n$ windows for each estimate.
Optionally accepts `opts`, a dict of optional arguments. All of these get
passed on to `us/seq-limit` to configure convergence checking.
See [[milne-sequence]] for more information about Milne's rule, caveats that
might apply when using this integration method and information on the optional
args in `opts` that customize this function's behavior."
:area-fn (comp first milne-sequence)
:seq-fn milne-sequence)