“…(Problem) φj strictly convex; linear equality (aj = 1); lj = 0 (Origin) Parallel subproblem in a RHS allocation algorithm for a problem with additional linear generalized upper bound (GUB) constraints (Methodology) Derived from the ranking algorithm from [Bod69]; replaces sorting by work with two heaps; no complexity analysis; also discusses the reoptimization of the sorted list, and discusses the invertibility of φ j and proposes a numerical approximation scheme (cites [Zip80b]) (Citations) Related work ([Bod69, LuG75, OhK80, Zip80b, BiH81, Roh79])-see [Roh82] for the published version of the latter; data structures ([Knu68, AHU74]) (Notes) The paper [Roh82] contains a discussion on the productivity of the activities, measured in terms of the quantity φj(x * j )/x * j , and relates this number to the value of φ j (0) for some special return functions given in [ChC58b,LuG75] [Bru84] P. Brucker, An O(n) algorithm for quadratic knapsack problems [Zie82]; initial interval given from the least/most costly items relative storage requirements, followed by a Newton/regula falsi step; proposes an algorithm based on the continued use of such steps, but without a formal convergence analysis > 0); lj = 0, uj = ∞ (Methodology) Improves the initial bounds from [VeK88]; cites [HMMS60,Lew81] for the origins of such bounds; also presents a Newton-type algorithm (denoted the "implicit algorithm") for obtaining µ * , which utilizes the bounding formula iteratively; the algorithm utilizes an initial ranking of inventory cost/storage requirement ratios, just as in [Zie82,VeK88]; convergence is claimed (referring to it being a Newton method) but not established [Gla96] P. Glasserman, Allocating production capacity among multiple products (Problem) φj(xj) = cj /γj (xj), γ −1 j convex; linear equality (aj = 1); lj > 0, uj = ∞ (Origin) Choice of base-stock levels and capacity allocations for a minimal total backorder or holding cost, in an inventory system with several items (Methodology) Simple heuristic decision rules (Citations) Similar analyses for other sequencing problems: [Kle76,Ana89]; optimization algorithms: [LuG75, Zip80b, IbK88] (Notes) Shows that simple rules exist (such as one that maximizes the time between stockouts) that behave asymptotically optimally, in the sense that as the number of orders tend to infinity the allocation policy tend to be optimal.…”