chenlin/Mujoco_WASM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add dense LU factorization and solve functions.

Browse Source 
PiperOrigin-RevId: 909290647
Change-Id: I77ae2352b20abf96ef7b48ae32b03a1f4098604e
This commit is contained in:
Yuval Tassa
2026-05-02 13:26:14 -07:00
committed by Copybara-Service
parent dbd451138c
commit 25751a7b98
3 changed files with 236 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/engine/engine_util_solve.h
+10
View File
@@ -94,6 +94,16 @@ MJAPI void mju_bandMulMatVec(mjtNum* res, const mjtNum* mat, const mjtNum* vec,
// address of diagonal element i in band-dense matrix representation
MJAPI int mju_bandDiag(int i, int ntotal, int nband, int ndense);
// dense LU factorization with partial pivoting
// factorizes n x n row-major matrix A in-place into L and U
// L has unit diagonal (not stored), U has explicit diagonal
// pivot stores row permutation: row i of original = row pivot[i] of result
// return 1 if successful, 0 if singular (diagonal element < mjMINVAL)
MJAPI int mju_factorLU(mjtNum* A, int n, int* pivot);
// solve A*x = b given LU factorization of A, LU and pivot are output of mju_factorLU
MJAPI void mju_solveLU(mjtNum* x, const mjtNum* LU, const mjtNum* b, const int* pivot, int n);
// sparse reverse-order LU factorization, assume tree topology (only dofs in index, if given)
// LU = L + U; original = (U+I) * L; scratch is size n
void mju_factorLUSparse(mjtNum *LU, int n, int* scratch,