Using Dynamic Programming Decomposition for Revenue Management with Opaque Products

Opaque products enable service providers to hide specific characteristics of their service fulfillment from the customer until after purchase. Prominent examples include internet-based service providers selling airline tickets without defining details, such as departure time or operating airline, until the booking has been made. Owing to the resulting flexibility in resource utilization, the traditional revenue management process needs to be modified. In this paper, we extend dynamic programming decomposition techniques widely used for traditional revenue management to develop an intuitive capacity control approach that allows for the incorporation of opaque products. In a simulation study, we show that the developed approach significantly outperforms other well-known capacity control approaches adapted to the opaque product setting. Based on the approach, we also provide computational examples of how the share of opaque products as well as the degree of opacity can influence the results.

Dynamic programming solutions for decentralized state-feedback LQG problems with communication delays

This paper presents explicit solutions for a class of decentralized LQG problems in which players communicate their states with delays. A method for decomposing the Bellman equation into a hierarchy of independent subproblems is introduced. Using this decomposition, all of the gains for the optimal controller are computed from the solution of a single algebraic Riccati equation. © 2012 AACC American Automatic Control Council).

Dynamic Programming Approaches for Estimating and Applying Large-scale Discrete Choice Models

People go through their life making all kinds of decisions, and some of these decisions affect their demand for transportation, for example, their choices of where to live and where to work, how and when to travel and which route to take. Transport related choices are typically time dependent and characterized by large number of alternatives that can be spatially correlated. This thesis deals with models that can be used to analyze and predict discrete choices in large-scale networks. The proposed models and methods are highly relevant for, but not limited to, transport applications. We model decisions as sequences of choices within the dynamic discrete choice framework, also known as parametric Markov decision processes. Such models are known to be difficult to estimate and to apply to make predictions because dynamic programming problems need to be solved in order to compute choice probabilities. In this thesis we show that it is possible to explore the network structure and the flexibility of dynamic programming so that the dynamic discrete choice modeling approach is not only useful to model time dependent choices, but also makes it easier to model large-scale static choices. The thesis consists of seven articles containing a number of models and methods for estimating, applying and testing large-scale discrete choice models. In the following we group the contributions under three themes: route choice modeling, large-scale multivariate extreme value (MEV) model estimation and nonlinear optimization algorithms. Five articles are related to route choice modeling. We propose different dynamic discrete choice models that allow paths to be correlated based on the MEV and mixed logit models. The resulting route choice models become expensive to estimate and we deal with this challenge by proposing innovative methods that allow to reduce the estimation cost. For example, we propose a decomposition method that not only opens up for possibility of mixing, but also speeds up the estimation for simple logit models, which has implications also for traffic simulation. Moreover, we compare the utility maximization and regret minimization decision rules, and we propose a misspecification test for logit-based route choice models. The second theme is related to the estimation of static discrete choice models with large choice sets. We establish that a class of MEV models can be reformulated as dynamic discrete choice models on the networks of correlation structures. These dynamic models can then be estimated quickly using dynamic programming techniques and an efficient nonlinear optimization algorithm. Finally, the third theme focuses on structured quasi-Newton techniques for estimating discrete choice models by maximum likelihood. We examine and adapt switching methods that can be easily integrated into usual optimization algorithms (line search and trust region) to accelerate the estimation process. The proposed dynamic discrete choice models and estimation methods can be used in various discrete choice applications. In the area of big data analytics, models that can deal with large choice sets and sequential choices are important. Our research can therefore be of interest in various demand analysis applications (predictive analytics) or can be integrated with optimization models (prescriptive analytics). Furthermore, our studies indicate the potential of dynamic programming techniques in this context, even for static models, which opens up a variety of future research directions.

Dynamic Programming Approaches for Estimating and Applying Large-scale Discrete Choice Models

People go through their life making all kinds of decisions, and some of these decisions affect their demand for transportation, for example, their choices of where to live and where to work, how and when to travel and which route to take. Transport related choices are typically time dependent and characterized by large number of alternatives that can be spatially correlated. This thesis deals with models that can be used to analyze and predict discrete choices in large-scale networks. The proposed models and methods are highly relevant for, but not limited to, transport applications. We model decisions as sequences of choices within the dynamic discrete choice framework, also known as parametric Markov decision processes. Such models are known to be difficult to estimate and to apply to make predictions because dynamic programming problems need to be solved in order to compute choice probabilities. In this thesis we show that it is possible to explore the network structure and the flexibility of dynamic programming so that the dynamic discrete choice modeling approach is not only useful to model time dependent choices, but also makes it easier to model large-scale static choices. The thesis consists of seven articles containing a number of models and methods for estimating, applying and testing large-scale discrete choice models. In the following we group the contributions under three themes: route choice modeling, large-scale multivariate extreme value (MEV) model estimation and nonlinear optimization algorithms. Five articles are related to route choice modeling. We propose different dynamic discrete choice models that allow paths to be correlated based on the MEV and mixed logit models. The resulting route choice models become expensive to estimate and we deal with this challenge by proposing innovative methods that allow to reduce the estimation cost. For example, we propose a decomposition method that not only opens up for possibility of mixing, but also speeds up the estimation for simple logit models, which has implications also for traffic simulation. Moreover, we compare the utility maximization and regret minimization decision rules, and we propose a misspecification test for logit-based route choice models. The second theme is related to the estimation of static discrete choice models with large choice sets. We establish that a class of MEV models can be reformulated as dynamic discrete choice models on the networks of correlation structures. These dynamic models can then be estimated quickly using dynamic programming techniques and an efficient nonlinear optimization algorithm. Finally, the third theme focuses on structured quasi-Newton techniques for estimating discrete choice models by maximum likelihood. We examine and adapt switching methods that can be easily integrated into usual optimization algorithms (line search and trust region) to accelerate the estimation process. The proposed dynamic discrete choice models and estimation methods can be used in various discrete choice applications. In the area of big data analytics, models that can deal with large choice sets and sequential choices are important. Our research can therefore be of interest in various demand analysis applications (predictive analytics) or can be integrated with optimization models (prescriptive analytics). Furthermore, our studies indicate the potential of dynamic programming techniques in this context, even for static models, which opens up a variety of future research directions.

Coordinated road-junction traffic control by dynamic programming

This paper presents a novel approach to road-traffic control for interconnected junctions. With a local fuzzy-logic controller (FLC) installed at each junction, a dynamic-programming (DP) technique is proposed to derive the green time for each phase in a traffic-light cycle. Coordination parameters from the adjacent junctions are also taken into consideration so that organized control is extended beyond a single junction. Instead of pursuing the absolute optimization of traffic delay, this study examines a practical approach to enable the simple implementation of coordination among junctions, while attempting to reduce delays, if possible. The simulation results show that the delay per vehicle can be substantially reduced, particularly when the traffic demand reaches the junction capacity. The implementation of this controller does not require complicated or demanding hardware, and such simplicity makes it a useful tool for offline studies or realtime control purposes.

Efficient articulated trajectory reconstruction using dynamic programming and filters

This paper considers the problem of reconstructing the motion of a 3D articulated tree from 2D point correspondences subject to some temporal prior. Hitherto, smooth motion has been encouraged using a trajectory basis, yielding a hard combinatorial problem with time complexity growing exponentially in the number of frames. Branch and bound strategies have previously attempted to curb this complexity whilst maintaining global optimality. However, they provide no guarantee of being more efficient than exhaustive search. Inspired by recent work which reconstructs general trajectories using compact high-pass filters, we develop a dynamic programming approach which scales linearly in the number of frames, leveraging the intrinsically local nature of filter interactions. Extension to affine projection enables reconstruction without estimating cameras.

Load profile based determination of distribution feeder configuration by dynamic programming

This paper presents a new method to determine feeder reconfiguration scheme considering variable load profile. The objective function consists of system losses, reliability costs and also switching costs. In order to achieve an optimal solution the proposed method compares these costs dynamically and determines when and how it is reasonable to have a switching operation. The proposed method divides a year into several equal time periods, then using particle swarm optimization (PSO), optimal candidate configurations for each period are obtained. System losses and customer interruption cost of each configuration during each period is also calculated. Then, considering switching cost from a configuration to another one, dynamic programming algorithm (DPA) is used to determine the annual reconfiguration scheme. Several test systems were used to validate the proposed method. The obtained results denote that to have an optimum solution it is necessary to compare operation costs dynamically.

A dynamic programming model for optimising feeding and slaughter decisions regarding fattening pigs

Costs of purchasing new piglets and of feeding them until slaughter are the main variable expenditures in pig fattening. They both depend on slaughter intensity, the nature of feeding patterns and the technological constraints of pig fattening, such as genotype. Therefore, it is of interest to examine the effect of production technology and changes in input and output prices on feeding and slaughter decisions. This study examines the problem by using a dynamic programming model that links genetic characteristics of a pig to feeding decisions and the timing of slaughter and takes into account how these jointly affect the quality-adjusted value of a carcass. The model simulates the growth mechanism of a pig under optional feeding and slaughter patterns and then solves the optimal feeding and slaughter decisions recursively. The state of nature and the genotype of a pig are known in the analysis. The main contribution of this study is the dynamic approach that explicitly takes into account carcass quality while simultaneously optimising feeding and slaughter decisions. The method maximises the internal rate of return to the capacity unit. Hence, the results can have vital impact on competitiveness of pig production, which is known to be quite capital-intensive. The results suggest that producer can significantly benefit from improvements in the pig's genotype, because they improve efficiency of pig production. The annual benefits from obtaining pigs of improved genotype can be more than €20 per capacity unit. The annual net benefits of animal breeding to pig farms can also be considerable. Animals of improved genotype can reach optimal slaughter maturity quicker and produce leaner meat than animals of poor genotype. In order to fully utilise the benefits of animal breeding, the producer must adjust feeding and slaughter patterns on the basis of genotype. The results suggest that the producer can benefit from flexible feeding technology. The flexible feeding technology segregates pigs into groups according to their weight, carcass leanness, genotype and sex and thereafter optimises feeding and slaughter decisions separately for these groups. Typically, such a technology provides incentives to feed piglets with protein-rich feed such that the genetic potential to produce leaner meat is fully utilised. When the pig approaches slaughter maturity, the share of protein-rich feed in the diet gradually decreases and the amount of energy-rich feed increases. Generally, the optimal slaughter weight is within the weight range that pays the highest price per kilogram of pig meat. The optimal feeding pattern and the optimal timing of slaughter depend on price ratios. Particularly, an increase in the price of pig meat provides incentives to increase the growth rates up to the pig's biological maximum by increasing the amount of energy in the feed. Price changes and changes in slaughter premium can also have large income effects. Key words: barley, carcass composition, dynamic programming, feeding, genotypes, lean, pig fattening, precision agriculture, productivity, slaughter weight, soybeans

Optimal Reservoir Operation for Flood Control Using Folded Dynamic Programming

Folded Dynamic Programming (FDP) is adopted for developing optimalnreservoir operation policies for flood control. It is applied to a case study of Hirakud Reservoir in Mahanadi basin, India with the objective of deriving optimal policy for flood control. The river flows down to Naraj, the head of delta where a major city is located and finally joins the Bay of Bengal. As Hirakud reservoir is on the upstream side of delta area in the basin, it plays an important role in alleviating the severity of the flood for this area. Data of 68 floods such as peaks of inflow hydrograph, peak of outflow from reservoir during each flood, peak of flow hydrograph at Naraj and d/s catchment contribution are utilized. The combinations of 51, 54, 57 thousand cumecs as peak inflow into reservoir and 25.5, 20, 14 thousand cumecs respectively as,peak d/s catchment contribution form the critical combinations for flood situation. It is observed that the combination of 57 thousand cumecs of inflow into reservoir and 14 thousand cumecs for d/s catchment contribution is the most critical among the critical combinations of flow series. The method proposed can be extended to similar situations for deriving reservoir operating policies for flood control.

Dynamic programming based optimum non-uniform samples for speech reconstruction and coding

Non-uniform sampling of a signal is formulated as an optimization problem which minimizes the reconstruction signal error. Dynamic programming (DP) has been used to solve this problem efficiently for a finite duration signal. Further, the optimum samples are quantized to realize a speech coder. The quantizer and the DP based optimum search for non-uniform samples (DP-NUS) can be combined in a closed-loop manner, which provides distinct advantage over the open-loop formulation. The DP-NUS formulation provides a useful control over the trade-off between bitrate and performance (reconstruction error). It is shown that 5-10 dB SNR improvement is possible using DP-NUS compared to extrema sampling approach. In addition, the close-loop DP-NUS gives a 4-5 dB improvement in reconstruction error.

Missile battery placement for air defense: A dynamic programming approach

A procedure to evaluate surface-to-air missile battery placement patterns for air defense is presented. A measure of defense effectiveness is defined as a function of kill probability of the defense missiles and the nature of the surrounding terrain features. The concept of cumulative danger index is used to select the best path for a penetrating hostile aircraft for any given pattern of placement. The aircraft is assumed to be intelligent and well-informed. The path is generated using a dynamic programming methodology. The software package so developed can be used off-line to choose the best among a number of possible battery placement patterns.