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
|
x = a*y + c
y = 4 * c
solve for [c] in terms of [a, x]
first solve for other unknowns: [y]
x = a*(4*c) + c
x = c * (1 + 4*a)
c = (1 + 4*a) / x
N equations in M unknowns, P constants
0 = x + y
0 = y*2 + 1 - x
-----------------------------------
solve_for(self, P: Vec<String>, M: Vec<String>, Q: HashMap<String, Expr>)
returns a ModelicaModel with fixed components; equations are LHS identifier, in
correct order, no extends, no connections, name suffixed " (rewritten)",
components as Parameter (w/ val), Input, or Output
V variables
Q constants (become kwargs)
P bound vars (independent, inputs/passed, become like constants)
M unknowns (dependent, outputs)
N' total equations
N equations with unknowns
V = Q + P + M
check: each Q is Integer or Float UnderSpecifiedConstant
filter N' to get N OverConstrained if extra equations
check: N >= M UnderConstrained
while len(eqns) > 0:
1. find first eqn with single unknown, and pop it
2. rebalance to get unknown on LHS
3. remove unknown from unknown list
return: sort equations to match order
solve_for(var) impl for SimpleEquation => Result<SimpleEquation, String>
check: var only on one side VariableNotFound or NaiveImplementation
put var on LHS
recursively apply rules to symmetrically unwrap LHS onto RHS
if try to unwrap Abs: NaiveImplementation
-----------------------------------
systems of differential equations:
will be N equations of N variables, in terms of implied time
should be no unknowns; odex thing is just varying initial equations
have JS boxes to show ranges of initial conditions
|
