The intersection–union definition of a submodular set function is
The diminishing returns characterization of submodularity says that its marginal increments decrease as the conditioning set grows:
To obtain diminishing returns from submodularity, apply the intersection–union inequality to and . Their intersection is and their union is , giving exactly the displayed comparison.
Conversely, assume diminishing returns, and enumerate . Put and . Since and , summing the marginal increment comparisons gives the telescoping sum
Rearrangement proves submodularity. When is empty, both sums are zero, so the proof includes that case.
Submodular rank function 2026-10-31
On a finite ground set , a submodular rank function is a normalised set function with , a monotone set function, and a submodular set function, such that for . Equivalently, require nonnegative integer values, monotonicity, submodularity and . The diminishing returns characterization of submodularity bounds each marginal by its singleton value. Thus .
These functions are precisely the rank functions of matroids. To see the converse direction, define . It contains the empty set. If , the unit marginal bounds give , so and heredity holds. For augmentation, suppose with . If every had zero marginal over , diminishing returns would keep each marginal zero as they are added successively. This would give , contradicting monotonicity. Some marginal is therefore one, giving an independent augmentation.
Finally any inclusion-maximal in this independence family has for every remaining . Successively adding them and using diminishing returns shows , so the resulting matroid really has rank . The forward direction is submodularity of matroid rank. These integer increment conditions distinguish matroid ranks from arbitrary real-valued submodular functions.