Suppose we are given a linear equation in $n$ dimensions:
$$Ax+By+Cz=S\,\,\,\,\,\, (n\text{ is }3)$$
with constraints on $x$, $y$ and $z$ as :
$$0\le x\le p,$$
$$ 0\le y\le q,$$
$$ 0\le z\le r.$$
We have to find the integral solution of this equation such that $x+y+z$ is minimum.
For example:
$$ 678 x+ 123 y+ 12 z=996$$
where
$$ 0\le x,$$
$$ 0\le y,$$
$$ 0\le z.$$
The solution which minimizes $x+y+z$ are $(1,2,6)$ and $(0,8,1)$.
What general methods can be applied to solve such a problem where we have constraint on all dimensions that they are greater than zero ?
This is the type of problems that can be solved by the Simplex algorithm combined by the Cutting plane method.
In general is hard to solve efficiently.
The feasible set (the set of points that satisfies the restrictions) lie on the portion of the plane $P=\{Ax+By+Cz=S\}$ that is inside the parallelogram defined by the inequalities. The intersection is a convex polygon. I think at most it is a hexagon but most of the time it has smaller number of sides.
The objective function $x+y+z$ gets smaller when you move in the direction of the vector $(-1,-1,-1)$. You can project this vector onto the plane $P$. That tells you in what direction to move inside the feasible set.
But you first need to find a feasible solution.