Nonlinear Programming: A Unified Approach

We consider a single-product revenue management problem where, given an initial inventory, the objective is to dynamically adjust prices over a finite sales horizon to maximize expected revenues. Realized demand is observed over time, but the underlying functional relationship between price and mean demand rate that governs these observations (otherwise known as the demand function or demand curve) is not known. We consider two instances of this problem: (i) a setting where the demand function is assumed to belong to a known parametric family with unknown parameter values; and (ii) a setting where the demand function is assumed to belong to a broad class of functions that need not admit any parametric representation. In each case we develop policies that learn the demand function “on the fly,” and optimize prices based on that. The performance of these algorithms is measured in terms of the regret: the revenue loss relative to the maximal revenues that can be extracted when the demand function is known prior to the start of the selling season. We derive lower bounds on the regret that hold for any admissible pricing policy, and then show that our proposed algorithms achieve a regret that is “close” to this lower bound. The magnitude of the regret can be interpreted as the economic value of prior knowledge on the demand function, manifested as the revenue loss due to model uncertainty.

/pdf/dynamic-pricing-without-knowing-the-demand-function-risk-3afky4ow1g.pdf

Dynamic Pricing Without Knowing the Demand Function: Risk Bounds and Near-Optimal Algorithms

We consider a single product revenue management problem where, given an initial inventory, the objective is to dynamically adjust prices over a finite sales horizon to maximize expected revenues. Realized demand is observed over time, but the underlying functional relationship between price and mean demand rate that governs these observations (otherwise known as the demand function or demand curve), is not known. We consider two instances of this problem: i.) a setting where the demand function is assumed to belong to a known parametric family with unknown parameter values; and ii.) a setting where the demand function is assumed to belong to a broad class of functions that need not admit any parametric representation. In each case we develop policies that learn the demand function "on the fly," and optimize prices based on that. The performance of these algorithms is measured in terms of the regret: the revenue loss relative to the maximal revenues that can be extracted when the demand function is known prior to the start of the selling season. We derive lower bounds on the regret that hold for any admissible pricing policy, and then show that our proposed algorithms achieve a regret that is "close" to this lower bound. The magnitude of the regret can be interpreted as the economic value of prior knowledge on the demand function; manifested as the revenue loss due to model uncertainty.

This paper systematically surveys the basic direction of development of stochastic quasigradient methods which allow one to solve optimization problems without calculating the precise values of objective and constraints function (all the more of their derivatives). For deterministic nonlinear optimization problems these methods can be regarded as methods of random search. For the stochastic programming problems, SQG methods generalize the well-known stochastic approximation method for unconstrained optimization of the expectation of random functions to problems involving general constraints.

Stochastic quasigradient methods and their application to system optimization

An apparatus and associated method for conditioning or reconditioning oil well drilling mud, comprising multiple hoppers for holding dry additive. Each hopper has a mechanism permitting the additive to be fed into an entrainment chamber at controlled, pre-selected rates. The additives are thus mixed at controlled rates with a free jet of mud forced across the entrainment chamber, to quickly achieve the desired density, viscosity and other properties of the mud. An augur is used to force the dry additive through an adjustable gate valve to control the additive flow rate. Or, the augur may be rotated at selected rates, and a fixed area additive flow passage employed between the hopper and the entrainment chamber.

E Sanvicente

Papers

Stochastic approximation of constrained systems with system and constraint noise