Convex Hull of Graphs of Polynomial Functions
5.2 Basic ideas of this chapter
(a) The graph of a polynomial function and a piece of the boundary (b) A subgraph on the boundary Figure 5.1:Investigating the boundary of the graph of a polynomial function
are presented, respectively. In this chapter, we focus on the characteristics of the convex hull of graphs of general polynomial functions over a polytope.
5.2 Basic ideas of this chapter
Consider the following example inR3. Let๐(๐ฅ, ๐ฆ) =๐ฅ2โ5๐ฅ๐ฆ+๐ฆ2be a polynomial function over the domain๐ ={(๐ฅ, ๐ฆ)โ[โ3,10]ร[โ3,10]}. For the constraint๐ง=๐ฅ2โ5๐ฅ๐ฆ+๐ฆ2, the feasible region, denoted by๐ฎ, is shown in Figure 5.1a which corresponds to the graph of๐ over domain๐. Recalling our MINLP algorithms, we add linear constraints to strengthen the LP relaxation. For any hyperplane๐ปinR3defined in the form{(๐ฅ, ๐ฆ, ๐ง)|๐ง=๐๐ฅ+๐๐ฆ+๐}
with constants๐, ๐, ๐โR,๐ปis said to be a linear underestimator to๐ over๐if ๐ฎ={(๐ฅ, ๐ฆ, ๐ง)|๐ง=๐(๐ฅ, ๐ฆ),(๐ฅ, ๐ฆ)โ๐} โ {(๐ฅ, ๐ฆ, ๐ง)|๐งโฅ๐๐ฅ+๐๐ฆ+๐}
โ โ
downward closed halfspace to๐ป
.
Graphically, it means that the corresponding downward closed halfspace completely contains the graph of๐ over๐.
In contrast to general MINLP algorithms, we want to find such linear underestimators directly.
They are expected to strengthen the LP relaxation. The intuition is that we only want to consider such hyperplanes๐ปthat support the graph, otherwise we can move it upwardly until the new generated hyperplane intersects the graph.
In other words, we say a linear underestimators๐ปisbelow(see Definition 5.16) the graph๐ฎ. Thus๐ปis said to bevalid.
(a) The view of the graph and a linear underestimator (b) The view from the other side Figure 5.2:A linear underestimator which supports two boundary points of the graph of a polynomial
function
To find linear underestimators๐ป, we study the intersection points๐ปโฉ๐ฎ. After a series of pre-liminary definitions in Section 5.3.1, we define locally and globally convex points in Section 5.3.2.
A point(๐ฅ0, ๐ฆ0, ๐ง0)on the graph is said to be locally convex if there exists๐ป โ (๐ฅ0, ๐ฆ0, ๐ง0) and๐ปis below the graph of๐ over a neighborhood of(๐ฅ0, ๐ฆ0, ๐ง0). A point(๐ฅ1, ๐ฆ1, ๐ง1)on the graph is said to be globally convex if there exists๐ป โ(๐ฅ1, ๐ฆ1, ๐ง1)and๐ปis below๐ฎ.
The hyperplane๐ป๐ก = {(๐ฅ, ๐ฆ, ๐ง) | ๐ง = 9๐ฅโ30๐ฆโ90}, shown as the yellow hyperplane in Figure 5.2, can be verified to be a linear underestimator for๐ over๐. The hyperplane ๐ป๐กintersects๐ฎ in two points(โ3,โ3,โ27)and(10,10,โ300). Hence(โ3,โ3,โ27)and (10,10,โ300)both are globally convex points. Note that they both are boundary points of๐ฎ. Consider further a point(๐ฅ0, ๐ฆ0, ๐ง0)such that the corresponding domain point(๐ฅ0, ๐ฆ0)is an interior point of๐. As we will show in Section 5.3.2, to check if(๐ฅ0, ๐ฆ0, ๐ง0)is globally convex, we need only to check if the corresponding tangent plane is below๐ฎ. However, in practice, it is quite hard to find those globally convex points such that the corresponding domain points are interior points of๐. In addition, the property of global convexity usually depends on the domain. On the one hand, any locally convex point may become globally convex if the domain size is small enough. On the other hand, a globally convex point with respect to the current domain could be only locally convex for a larger domain. Note that in the example above๐ is neither convex nor concave over๐.
Now we move our attention to those globally convex points for which the corresponding domain points are on the boundary of๐. Consider the subgraph with restriction๐ฆ = โ3, which is presented as
{(๐ฅ, ๐ฆ, ๐ง)|๐ง=๐(๐ฅ, ๐ฆ),โ3โค๐ฅโค10, ๐ฆ=โ3}.
5.2 Basic ideas of this chapter
(a) Subtangent plane (b) Globally convex boundary point inR2 Figure 5.3:Example for a globally convex boundary point
This subgraph is shown as the red curve in Figure 5.1a. Since๐ฆ =โ3is satisfied for any point in the red subgraph, after projecting the space{(๐ฅ,โ3, ๐ง)} โR3to the space{(๐ฅ, ๐ง)} โR2, we get an isomorphic two-dimensional curve inR2
{(๐ฅ, ๐ง)|๐ง=๐(๐ฅ,โ3) =๐ฅ2+ 15๐ฅ+ 9 =: ห๐(๐ฅ),โ3โค๐ฅโค10}.
In general, we show at the beginning of Section 5.3.3 that certain subgraphs on the boundary can be projected to an isomorphic graph in a space with lower dimension. The one-dimensional curve is shown as the red curve in Figure 5.1b. Note that the corresponding function๐หis a univariate function. Fortunately, the study of the convexity of univariate functions is much easier than that for bivariate functions. In the example๐หhas domain[โ3,10]and is a convex function over[โ3,10].
According to the definition of globally convex points, any point๐ฅ* โ[โ3,10]in Figure 5.1b is globally convex in the projected spaceR2. Theorem 5.12 implies that for any such๐ฅ*, because ๐ฅ*is globally convex in the projected space, the boundary point(๐ฅ*,โ3)in๐is also globally convex in the original space. This means there exists a hyperplane๐ป โ(๐ฅ*,โ3, ๐(๐ฅ*,โ3))and ๐ปis below๐ฎ. Consider the case๐ฅ*= 0. Then(0,โ3,9)is a globally convex point. Figure 5.3b shows that in the projected spaceR2, the tangent plane, shown as the green line, is the unique underestimator. The corresponding line inR3, shown as the green line in Figure 5.3a, is then {(๐ฅ,โ3,15๐ฅ+ 9)|๐ฅโR}which is defined as subtangent plane in Section 5.3.3. Corollary 5.13 implies that every linear underestimator๐ปwith๐ปโ(0,โ3,9)satisfies
๐ปโ {(๐ฅ,โ3,15๐ฅ+ 9)|๐ฅโR},
which means any valid hyperplane which contains(0,โ3,9)always contains the green line.
The blue line{(๐ฅ,โ3,9๐ฅ)|๐ฅโR}in Figure 5.2a is a subtangent plane on(โ3,โ3,โ27), as defined in Section 5.3.3. We can verify that the yellow hyperplane is the affine hull of the blue line and the point(10,10,โ300), i.e.,
๐ป๐ก={(๐ฅ, ๐ฆ, ๐ง)|๐ง= 9๐ฅโ30๐ฆโ90}= aff{{(๐ฅ,โ3,9๐ฅ)|๐ฅโR},{(10,10,โ300)}}. For any point(10,10, ๐ง1)with๐ง1<โ300, we can also verify that
๐ป๐= aff{๏ธ{(๐ฅ,โ3,9๐ฅ)|๐ฅโR},{(10,10, ๐ง1)}}๏ธ
is also a linear underestimator. By comparing๐ป๐กand๐ป๐we have
๐ป๐กโฉ ๐ฎ ={(โ3,โ3,โ27),(10,10,โ300)}){(โ3,โ3,โ27)}=๐ป๐โฉ ๐ฎ.
From the intuition, we prefer๐ป๐กsince the resulting relaxation is tighter. For this purpose we define tight and loose hyperplanes in Section 5.3.4. In general, a valid hyperplane๐ป๐ is definitely loose if there exists another valid hyperplane๐ป๐กwhich preserves all intersection points and intersects in additional point(s) with๐ฎ, which means
(๐ป๐กโฉ ๐ฎ))(๐ป๐โฉ ๐ฎ).
This is a sufficient but not necessary condition for loose hyperplanes. Using Lemma 5.26 in Section 5.3.4 we verify that the yellow hyperplane in Figure 5.2a is a tight hyperplane.
After that, in Section 5.3.5 we prove for every loose hyperplane๐ป๐that there exists a tight hyperplane๐ป๐กthat preserves intersection points with
(๐ป๐กโฉ ๐ฎ)โ(๐ป๐โฉ ๐ฎ).
We call the corresponding halfspaces tight or loose halfspaces. Note that in the example above we have๐ป๐กโฉ๐ป๐ ={(๐ฅ,โ3,9๐ฅ) |๐ฅโR}which is the blue line in Figure 5.2a. Graphically, we can rotate๐ป๐around the blue line as axis to generate๐ป๐ก. The rotation approach is the basic idea of a few proofs in this section.
Finally, in Section 5.3.6, we prove that to form the convex hull of๐ฎusing halfspaces, we only need tight hyperplanes. In other words, any loose hyperplane is proved to be redundant.
In Section 5.3 we only include theoretical results. We cannot use them to solve MINLP directly. In Section 5.4 we develop algorithms to compute tight hyperplanes for the graph of bivariate polynomial functions with degree up to3over a polygon in R2. Note that the domain does not have to be box-constrained. In the algorithms, we first find all globally convex domain points on the boundary. This is very tractable since we only need to find globally convex points in the graph of univariate polynomial functions with degree3over a closed interval inR. Based on those globally convex domain points, the algorithms find a series of tight halfspaces. Computations in Section 5.5 show that these tight halfspaces improve the dual bounds significantly.
5.3 Convex hull of graphs of polynomial functions over a polytope