“…Following the same logic as in [3], it is required that α k , αk ≥ 0 for all k,k, which implies that M ± f ≥ 0, so the resulting computational matrix A(h) is an M-matrix and the method preserves the solution positivity. If this is not the case, then other equations of form (22), (24), (25), or transformed equations such as (27), are chosen to form (29).…”