Efficient Resource Allocation with Non-Concave Objective Functions

We consider resource allocation with separable objective functions defined over subranges of the integers. While it is well known that (the maximization version of) this problem can be solved efficiently if the objective functions are concave, the general problem of resource allocation with non-concave functions is difficult. In this article we show that for fairly well-shaped non-concave objective functions, the optimal solution can be computed efficiently. Our main enabling ingredient is an algorithm for aggregating two objective functions, where the cost depends on the complexity of the two involved functions. As a measure of complexity of a function, we use the number of subintervals that are convex or concave.