A Partial Ranking Algorithm for Resource Allocation Problems

We present an algorithm to solve resource allocation problems with a single resource, a convex separable objective function, a convex separable resource-usage constraint and bounded variables.Through evaluation of specific functions in the lower and/or upper bounds, we obtain information on whether or not these bounds are binding.Once this information is available for all variables, the optimum is found through determination of the unique root of a strictly decreasing function.A comparison is made with the currently known most efficient algorithms.