Nonnegative diagonal P of the objective, one coefficient per variable; zero is allowed.
OptionalinitialOptional complete variable vector admitted like linear. The current kernel ignores it and starts cold; omission never reuses a previous solution.
Linear q coefficients with the same variable order and length as diagonal.
Inclusive sparse bounds on A*x in caller-defined row units; an empty list means no row constraints.
Borrowed program minimizing 0.5*sum(diagonal[i]*x[i]^2)+sum(linear[i]*x[i]) under inclusive sparse row bounds. Coefficient units follow the caller's variable, objective and row units; no geometric conversion occurs.
admitAutoMovieQuadraticProgram requires 1..2^31-1 variables and at most 2^31-1 rows, complete aligned vectors, nonnegative diagonal, unique resident row indices, ordered bounds and finite scalars with magnitude below 1e30. Container shape and allocatable aggregate storage remain caller/host premises. Admission does not establish feasibility or convergence.