With the advances in technologies and the growing popularity of e-commerce, huge datasets and massive computational power have never been more accessible. Moreover, the rising machine learning and distributionally robust optimization techniques bring new opportunity for more effective inventory decision making in data-rich environments. The goal of this dissertation is, thus, to explore data-driven approaches to inventory management problems that are efficient and practical.

We address the challenges in this field from three different aspects: firstly, we aim at proposing a flexible model for capturing real-world demand process accurately with as few assumptions as possible; then, we take into account additional features in the demand model and derive robust inventory policies with out-of-sample performance guarantees under milder assumption than current literature; and finally, we explore the usage of a group of decomposition algorithms to tackle the increasing computational difficulty as the data size grows. Chapter 2, Chapter 3 and Chapter 4 each delves into one of these three directions respectively.

In Chapter 2, we leverage the universal approximating capability of neural network structures to approximate an arbitrarily complex autoregressive demand process without any parametric assumptions. By adopting a quantile loss in training, we allow our neural network to output directly an estimation of the critical quantile, which is indeed the inventory policy for classical newsvendor problem. In addition, in contrast to the prevalent feedforward neural networks which are asymptotically stationary, the special structure we choose is capable of handling nonstationary time series. To the best of our knowledge, this is the first approach which deals with nonstationary time series without any parametric assumption or preprocessing to capture the components like trend or seasonality. Though theoretical guarantees are sacrificed due to a lack of assumption on the underlying real process, empirical studies validate the performance of our approach on real-world nonstationary demand process. Moreover, we establish the optimality of the myopic policy to the multi-period newsvendor problem where unmet demand and excess inventory can be carried over to the next period and argue that our approach is also a data-driven solution.

The second project in Chapter 3 addresses the data-driven newsvendor problem from a different angle with the goal to achieve robust policies as well as theoretical support. We start with a simple linear demand model to incorporate information from other features related to the demand, such as price, materials and etc. And to hedge against the uncertainty of the demand distribution, the idea of distributionally robust optimization (DRO) is applied. We contribute to the current literature of DRO applications in supervised learning by adopting a fixed design interpretation of the features. Thus, similar to the neural network approach, we are also able to relax the assumption of identical and independent sample points, which is more applicable in real-world scenarios. Then, by leveraging results from fixed design linear regression, we propose a two-step framework to obtain a newsvendor solution. Moreover, the Wasserstein metric is chosen for constructing the ambiguity set of all candidate distributions and based on which our data-driven policy can be obtained efficiently in polynomial time and attains both finite-sample and asymptotic performance guarantees.

Finally, in Chapter 4, we put some effort into dealing with practical issues of implementing such data-driven approaches when massive datasets are available. Specifically, we consider a group of decomposition algorithms which are suitable for large-scale multi-block convex optimization problems with linear constraints. This problem setting covers a lot of applications in machine learning and data-driven problems. We focus on a special case of such algorithms which can be guaranteed to converge under mild conditions with linear rate, and also enjoys the convenience of parallel implementable subproblems. We modify an adaptive parameter tuning scheme to achieve faster convergence in practice. And at the end, we further show that when parameters are chosen appropriately, global convergence can be established even if the primal subproblems are only solved approximately.

We reckon that our results are just some primary attempts at achieving the goal of efficient decision making in a data-driven environment, and hope that this dissertation can serve as a catalyst for other research in this field. Thus, we list a number of directions for future research in the last chapter after the concluding remarks.

Traditionally, stochastic models in operations research use specific probabilistic assumptions to model random phenomena, and determine optimal policies or decisions on this basis. Often, these probabilistic assumptions are parametric, and entail estimation of parameters using very small samples of data. Many a times, the available information is not sufficient to postulate a model with any degree of certainty. Consequently, policies based on parametric assumptions in this case, are very sensitive to the particular assumptions made. One of the goals of this thesis is therefore the development of objective, adaptive, data-driven, learning approaches to objective functions, that make as few parametric assumptions as possible, and give rise to optimal policies that perform well for small samples, without compromising large sample performance. While this clearly seems a very difficult problem, it is one that is observed in nearly every operations management problem and is certainly the right problem to pursue. In this thesis, we develop novel learning approaches to specific problems in inventory control, call center staffing and dynamic assortment optimization. We test these approaches computationally, and provide strong evidence for the adoption of our general approach in tackling model uncertainty in operations management problems.

The destructive potential of wildfires has been exacerbated by climate change, causing their frequencies and intensities to continuously increase globally. In this context, increasing wildfire activity across the globe has become an urgent issue with enormous ecological and social impacts. Wildfires have consumed important areas and forest resources, as a result, fire management expenditures have increased and thousands of homes and many lives have been lost. Moreover, they have significantly impacted biodiversity and greenhouse gas emissions on a global scale.

The current incidents across the globe highlight the need for preemptive policy measures to reduce the risk of fire occurrence, managing the land in an effective way to protect natural forests, agricultural areas, and human lives. These concepts are included in what is known as FireSmart Forest Management (FSFM). This paradigm considers opportunities in three dimensions: i) decrease of the fire behavior potential of the landscape, ii) reduction of the potential for fire ignitions, and iii) increase in the fire suppression capability.

This dissertation aims at advancing the theory, practice, and large-scale implementation of complex data-driven decision making and machine learning models in the context of landscape management under wildfire risk, integrating Operations Research, Computer Science, and Data Science techniques. We focus our efforts on the understanding, evaluation, and development of effective prevention and mitigation policies, with the potential of being implemented practice, as well as exploring and developing new FSFM techniques.

We divided our study into three main aspects: Simulation, Decision-Making, and Machine Learning. In Chapter 1, we focus on the development and evaluation of an accurate, flexible, and efficient wildfire simulation model that can be integrated with data-driven decision-making models. Empirical results on thousands of simulations show the high performance of the model compared to existing solutions, highlighting its accuracy with real-life instances. We then focus our efforts on its generalization in Chapter 2, seeking to adjust its main parameters to mimic the fire spread behavior observed in different regions of the world where no empirical models are available. This, exploiting historical information for training purposes using derivative-free optimization techniques to adjust the parameters of the model, allowing us to capture current wildfire dynamics. Experiments performed on datasets located in different regions of the world show the potential of the proposed method.

Second, in Chapters 3 and 4, we explore the integration of this model with landscape planning decision-making models to derive robust fuel treatment policies to mitigate expected losses due to future wildfire events, generating fire resilient landscapes. We study and compare complex network algorithms to develop a mathematical model denoted Downstream Protection Value, capturing the importance of the different components of the land to provide a natural prioritization of where mitigation actions should be implemented. An optimization framework incorporating multiple variables to analyze the inherent trade-offs involved in the planning process is developed, providing practitioners and researchers with an open-source decision support system implementation involving multiple and potentially, opposing objectives. We evaluate the performance of the proposed mathematical model compared to existing solutions, highlighting its superior performance with thousands of experiments involving uncertainty for landscapes located in North America. Several extensions are discussed providing future research directions in the field. We further expand this framework in Chapter 5, incorporating wildfire suppression strategies derived from a novel multi-agent decision-making model. In this application, a group of agents is deployed to the field once an ignition or active fire is detected with the aim of containing or stopping it as soon as possible. The sequential and temporal dimensions of the problem become a challenge to apply traditional modeling techniques. We develop a deep reinforcement learning algorithm focused on exploiting the collaboration and coordination between independent agents. Extensive computational results demonstrate the impact of including local rewards and the concept of sub-groups of agents in the context of centralized training and decentralized execution algorithms, leading to more complex and effective collaboration strategies between agents belonging to the same group and the environment in general.

As part of our decision support system extensions, we build end-to-end machine learning models to understand the wildfire phenomenon from a large-scale perspective. For this, in Chapter 6, we explore the impact of different landscape compositions on the risk of wildfire ignition using remote sensing data to support challenging landscape planning decisions. Using a custom convolutional neural network model integrated with state-of-the-art visualization techniques, we highlight the main areas of interest for the deep learning model, focusing our efforts on the interpretability of the results. This, with the aim of opening the artificial intelligence black-box to fully understand the rationale behind the results and the different risk levels associated with characteristic spatial patterns observed in the land. We validate our results with previous studies using similar datasets, noting how the proposed model significantly surpasses their predictive performance while providing insights about the learning process of the model.

Finally, in Chapter 7, we develop a global study of the main characteristics and drivers of wildfire regimes consolidating observations for almost two decades of wildfires. We classify and delineate regions of the world sharing similar fire activity as well as identifying their driving factors to support national or regional wildfire prevention/mitigation policies using a variety of machine learning techniques. To the best of our knowledge, this represents the first study that defines fire regimes spatially at a global scale bridging existing knowledge gaps between global and regional fire studies.

Our results represent an attempt at improving the integration of multiple disciplines in the context of effective data-driven decision-making under natural hazards uncertainty. Several challenges are still open. We hope that this research can serve as a motivation to expand the field's perspective with high-impact applied projects involving mathematical, ecological, economic, data, and social sciences.

Online recommendation systems ask these questions everyday: How to describe customers' purchasing behavior? How to design a product assortment in order to maximize their expected profit/revenue with given customers' behavior? What is the optimal pricing strategy for an assortment? Even further, how to jointly design optimal assortment and pricing at the same time in an efficient way? The answers to these questions have a direct influence to the profitability and feasibility of the recommendation systems.

My thesis handles assortment and price optimization problems with various applications in online e-commerce and travel related recommendations, such as flight, rental car and hotels. For example, for car rentals in Expedia, they would like to offer customers a recommendation page of cars with different brands, prices, types, options, etc. What is a good recommendation for the customers so that they would have a good shopping and traveling experiences? First, it needs to be relevant to customers' choice behavior that can be learned from previous purchasing history or from marketing surveys; second, the recommendation cannot be too specified, which means that those rental cars in the recommendation page cannot be too similar in terms of their attributes; third, it cannot take too long to show the recommendations to the customers - an efficient algorithm is required. In this thesis, we will show our approach to assortment and price optimization problems.

The main contributions of my thesis is: 1) We formulate the assortment and price optimization problems in a choice-based way, which provides a good balance between relevance and variance of the products in an assortment; 2) We develop applicable recommendation algorithms that run in polynomial time and can be dynamically adapted; 3) Compared to the previous literature, our results are more advanced in terms of efficiency and applicability. Specifically, this thesis is consist of three essays in choice-based assortment and price optimization problems.

In the first essay, we study the joint constrained assortment and price optimization problem under the nested logit model with a no-purchase option in every choice stage. The cardinality or space constraints are imposed separately on the assortment of products that are offered in each nest. Specifically, cardinality constraint on a nest limits the total number of products that can be offered in that nest, and space constraint on a nest limits the total space consumption of products within that nest. The goal is to jointly determine the optimal assortment with optimal prices to maximize the expected profit per customer under cardinality or space constraints. By using our solution approach, this problem is simplified to find the fixed point of a single-variable unimodal expected profit function, where efficient searching algorithms can be applied. Furthermore, we provide a piecewise convex fixed point representation to facilitate computing. The optimal solution under cardinality constraints and a 2-approximate solution under space constraints can be obtained efficiently.

In the second essay, we study choice-based constrained assortment and price optimization problems under the multilevel nested logit model with a no-purchase option in every choice stage. For the constrained assortment optimization problem, each candidate product is associated with a fixed profit. The goal is to identify the optimal assortment satisfying cardinality or space constraints to maximize the expected profit per customer. Under cardinality constraints, there is a limitation imposed on nodes in the second lowest level. A polynomial-time algorithm with computational complexity $O(n\max\{m, k\})$ is provided to locate the optimal assortment for the $m$-level nested logit model with $n$ products, where $k$ is the maximum number of products within any node in level $m-1$. Under space constraints, every product consumes a certain amount of space and candidate assortments must satisfy the space limitation. However, the assortment optimization problem becomes NP-hard under space constraints, thus we develop an algorithm to find a 2-approximate solution in $O(mnk)$ operations. For the price optimization problem, we aim to find the profit-maximizing prices for all products. With product-differentiated price sensitivities, the expected profit function is no longer concave even under the two-level nested logit model, but we are able to reduce the multiproduct price optimization problem from a high dimensional optimization to the maximization of a unimodal function in single-dimensional searching space, in which the optimal prices can be found in a tractable manner.

In the third essay, we know that assortment and pricing decisions are of significant importance to firms and have huge influences on profit. How to jointly optimize over both assortment and prices draws increasing attention recently. However, in the most existing literature that considers joint optimization problem, they either impose strong restrictions on the choice structure or have strong assumptions on the price sensitivity parameters. Moreover, currently there is no flexible and comprehensive way to deal with the joint effect of assortment and pricing under multistage choice structure since the tangle between the assortment and prices makes the joint optimization problem less tractable. In this paper, we study the joint capacitated assortment and price optimization problem where the consumer choosing behavior is governed by the multistage tree logit model. Under the cardinality constraints, we develop an efficient algorithm that runs in polynomial time to find the optimal assortment with optimal prices. Under the space constraints, the assortment optimization problem is NP-hard even under tree logit model with only two levels. We can obtain a 2-approximate solution within the same time scale compared to the joint optimization problem under cardinality constraints. For a tree logit model with $N$ candidate products, both algorithms run in $O(GN\log G)$ where $G$ is the number of grid points for each node. The complexity can be further reduced to $O(GN\log K)$ under mild conditions, where $K$ is the maximum number of children nodes that a nonleaf node could have in the tree structure.

This dissertation studies the data-driven approaches to flexible systems design problems under uncertainty. We discuss real applications in various contexts with flexibility: the capability to satisfy different types of customer demands (e.g. one-way and round trips in the context of car sharing systems); the geographical demand distribution estimation and associated inventory allocation; and the freedom in production plans to fulfill uncertain customer demands (e.g. flexible recipes in continuous production process).

The problems we consider have different objectives and more importantly several degrees of richness in data availability. We develop data-driven optimization models accordingly. Specifically, in the case of new market expansion for example, the firm has to make one-shot decision with limited or side information. The focus of data-driven approach in this case is on the portability of information. Distributionally-robust optimization methodologies are applied to derive strategic decisions that hedge the risks. At the tactical level, e.g. resource planning, the firm deploys planning with ample historical data. For online retailers, geographical demand distributions need to be estimated from historical sales and serve as key input to their regular inventory allocation decisions. Furthermore, operational decisions generally require more detailed data, especially the continuous data for real-time decisions. We study the problem where routine production plans are chosen together with raw material investment decisions when periodic demand data may be available.

In the first part of the dissertation, we study the planning problem faced by urban electric vehicle (EV) sharing systems, that offer both one-way and round trips, in designing the geographical service region. This decision encompasses the trade-off between maximizing customer adoption by covering travel needs, and controlling fleet operations costs. We develop a mathematical programming model that incorporates details of both customer adoption behavior and fleet management (including EV repositioning and charging) operations under spatially-imbalanced and time-varying travel patterns. To address uncertainty in customer adoption, we employ a distributionally-robust optimization framework that informs robust decisions to avoid possible ambiguity (or lack) of data. Mathematically, the problem is approximated by a mixed integer second-order cone program (MISOCP), which is computationally-tractable. Applying this approach to the case of Car2Go's service in San Diego, California, with real operations data, we investigate several planning questions and suggest potential for future development of the service.

To make better inventory allocation to distribution centers, understanding of the geographical demand distribution is essential to online retailers who possess historical sales data that might be contaminated and/or with missing data. The second part of the dissertation presents two models: the first model estimates the geographical demand distribution; the second model integrates the demand estimation together with inventory optimization. In the first model, we study the missing geo-demand data completion problem for a national online retailer. We formulate the problem as a low-rank tensor recover problem in a convex optimization framework. An alternating direction augmented Lagrangian (ADAL) method has been developed and tailored for solving the tensor recovery problem with partial observations. We first discuss efficiency and effectiveness of the algorithm via experiments with synthetic data. We then apply the framework with observed geo-demand from the online retailer. Finally, the benefits of the missing geo-demand data completion are summarized based on computational experiment results. We have shown that the recovered geo-demand distributions possesses more smoothness over time and rendered better generalization performance than the observed geo-demand upon integrated into the existing learning framework. We also integrate the missing data recovery with the data-driven newsvendor model which provides estimation of demands as well as optimal order quantity. A preliminary analysis shows that the proposed model preserves the condition for optimal order quantity as it is in the data-driven newsvendor model. Future work directions are also discussed.

The last part of this dissertation focuses on the inventory investment, recipe selection and resource allocation decisions in continuous process systems with flexible recipes under demand uncertainty. Due to variations in both raw material quality and market conditions, variations in the recipes are used in continuous production processes. Such flexibility is not on design but on the operation that allows adjustments of recipe items aiming to achieve better input utilization than traditionally fixed recipes. We develop a two-stage stochastic mixed integer program formulation and propose a heuristic to the second stage allocation optimization problem. In the first stage, the model determines inventory levels for each period based on past demand data. After demand arrivals are realized, the second stage recourse makes recipe selection and allocation decisions in production. With available historical demand data, a simulation-based approach based on SAA algorithm is developed to solve the stochastic program. The results of numerical study show the performance of the approach on various cost settings as well as the benefits of flexible recipes over fixed recipes. In the proposed approach, we focus on the application of the sample average approximation (SAA) algorithm and use Bootstrap sampling as the default in demand simulation. A direction of future improvement is to incorporate better techniques in the simulation of future demand arrivals based on historical demand data. Those techniques may consider some properties of the demand, such as seasonality and autocorrelation. Also, with limited demand information, a robust optimization model might be developed that considers the worst cases. Moreover, since our model assumes any inventory leftover at the end of each period is disposed, the extension that relaxes this assumption and introduces inventory holding cost in multi-period setting should also be investigated.

Disasters strike suddenly and cause destruction which disrupts the availability of basic survival supplies for people living in affected areas. The efficiency of humanitarian organizations in providing relief has a direct and crucial impact on the survival, health, and recovery of affected people and their communities. To better prepare to respond to disasters, many relief organizations use supply pre-positioning. However, the real and potential needs of different locations change over time and when an organization uses traditional warehouse pre-positioning, relief operations are limited by set inventory locations that are difficult to alter. For this reason, a well known organization recently considered including a large supply holding ship in its operations. By holding inventory on a ship, the organization would be able to dynamically relocate its inventory over time in response to changing relief supply demand forecasts.

To our knowledge, the research contained herein is the first to examine dynamic inventory relocation for responding to disasters over time. Specifically, we examine how to optimally relocate and manage inventory for a single mobile inventory to serve stochastic demand at a number of potential disaster sites over time. While we keep in mind the motivating example of a supply holding ship in the disaster relief setting throughout this dissertation, the model and most of the results are applicable to any type of mobile inventory, facility, or server in any setting.

We first examine the dynamic relocation problem. We model the problem using dynamic programming and develop analytical and numerical results regarding optimal relocation policies, the optimal path and speed of relocation decisions, and the value of inventory mobility over traditional warehouse pre-positioning. To help overcome the computational complexity of the problem, we develop a heuristic which solves relatively large problem instances in our numerical experiments within 0.5% of optimality in less than 0.1% of the time required by an exact algorithm.

As it is suboptimal to consider relocation decisions and inventory management decisions separately, we also examine the joint dynamic relocation and inventory management problem. To our knowledge, we are the first to examine the dynamic relocation and inventory management problem with stochastic demand. Similarly to the dynamic relocation problem, we model this problem using dynamic programming. We develop a number of analytical results characterizing the optimal relocation and inventory management policies.

As the first to examine these problems, we hope this research serves as a catalyst for other research in this area; accordingly, we conclude this dissertation by discussing a number of areas for future research.

This dissertation contributes to the supply chain design literature by providing a novel solution approach and presenting new theoretical and applied models. Using the proposed solution approach, we are able to formulate models that consider joint facility location and inventory management problems in a stochastic environment. Due to these uncertainties, it is important to design the supply chain that better adapts to unexpected changes.

In Chapter 2, we study several joint location-inventory problems. In particular, we consider cases with uncapacitated facilities, capacitated facilities, correlated retailer demand, stochastic lead times, and multi-commodities. We show how to formulate these problems as conic quadratic mixed-integer problems. We compare with the existing modeling and solution methods to state that the new conic integer programming approach not only provides a more general modeling framework but also leads to fast solution times in general. Valid inequalities, including extended polymatroid and extended cover cuts, are added to strengthen the formulations and improve the computational results.

In Chapter 3, we employ the technique presented in the previous chapter to explore the supply chain design problem that incorporates reverse logistics decisions. Supply chains with returned products are receiving increasing attention in the operations management community and this chapter studies a capacitated facility location model with bidirectional flows and marginal value of time for returned products. While at the retailers' site, products can be shipped back to the supplier for reprocessing. Each distribution center is capacitated and handles stocks of new and/or returned product.

In Chapter 4, we study another applied area that is getting increasing attention in the operations management community: the nonprofit sector. Under some circumstances nonprofit practices can significantly differ from for-profit ones and this might require a different design of the supply chain. This chapter starts with a definition of the nonprofit supply chain identifying its major managerial challenges. Then, we provide a summarized comprehensive overview of current operations management literature that address each major challenge. The second part of the chapter studies a particular nonprofit supply chain design problem in the context of a humanitarian organization distributing food in a less secure area.

My thesis research explicitly emphasizes integrating marketing and operations management (hereafter OM) perspectives in the formulation of strategy. Shortened product life cycles, technological advancements in products and processes, globalization of markets, consumerism, and the rapidity of change have only exacerbated the perceived need to link Marketing/OM strategies. My primary research interest follows the trend and explores the science of delivering value to customers from an integrated view of marketing and OM with special emphasis on the timing of introducing product lines (Chapter 1) and the role of social contagion in forming customer lifetime value (Chapters 2 and 3).

In Chapter 1, we study the problem of when to introduce a line extension of a product with an existing version in an integrated inventory (supply) and diffusion (demand) framework. The launch of a new product with successive (and differentiated) versions always commands a large commitment of resources in production and marketing, thus the introduction strategy often requires careful planning. A key element in the introduction strategy is the introduction time. There is yet a formal model to quantify the impact of inventory cost on product line introduction timing decisions considering the demand dynamics in product life cycle and substitution among versions. This paper takes a first step towards filling this gap. On the demand side, we consider the demand dynamics of both versions during product life cycle, in marketplaces where repeated industry practices are observable to customers. Based on the Bass model, we propose a splitting Bass-like diffusion model to describe the adoption processes for the two successive (and differentiated) versions of one product, taking into account the role of customer expectation in shaping purchase choices. On the supply side, we model the impact of inventory holding cost that arises from a simple ordering policy. We show there exists a unique optimal time to introduce the line extension in the planning horizon. We quantify the optimal launch-time and both versions' sales trajectories. In contrary to the existing optimal policy in the literature (i.e., ``Now or Never''), we find that the optimal introduction can happen anytime from ``Now'' to ``Never'', depending upon the characteristics of different products. We show that when inventory holding cost is small and the ordering cycle is short, the optimal introduction time is indeed ``Now '' or ``Never''. However, as inventory holding becomes substantial, the firm might choose to delay the introduction when the line extension is more profitable than the existing version, or to accelerate the introduction when the existing version generates more profit. Our integrated model sheds light on the necessity of coordinating marketing and operations management decisions.

In Chapters 2 and 3, we incorporate social contagion into customer lifetime value analysis. Prior research has assumed that a customer's lifetime value (LV) only depends on her own purchase history. The rise of Internet and viral marketing casts doubt on this assumption. In the Web 2.0 economy, social contagion is so integral to customer's shopping process that purchase behaviors are frequently interdependent. We investigate how social contagion might influence a customer's lifetime value beyond her own purchases. We posit that a customer's total lifetime value (LV) is a sum of her total purchase value (PV) (accounting for others' influence on her purchases) and her total influence value (IV). Specifically we have: LV = PV + IV. Consequently a customer can still have a high lifetime value even if she has a low PV as long as she has a high IV.

Chapter 2 presents a model with homogeneous population. Building on the classical Bass diffusion model, we show that PV, IV, and LV decrease in the convex manner with adoption time. Hence a customer who adopts earlier is much more valuable than a customer who adopts later. While PV increases with the innovation parameter, IV decreases with it. Early adopters have their LV decrease with innovation parameter while later adopters have their LV increase with it. Interestingly, PV decreases with the imitation parameter and IV increases with it for early adopters and decreases with it for late adopters. LV increases with the imitation parameter if the timing of adoption is below a cutoff value and decreases with it if it is above the cutoff. We then examine how a firm might improve its overall customer LV by accelerating purchase made possible by offering introductory price discounts to a subset of customers. We characterize the optimal size of the targeted customers in terms of level of discount, innovation as well as imitation parameters and demonstrate that the firm can significantly increase its total customer LV by purchase acceleration. We also analyze the impact of purchase deceleration in a make-to-stock supply chain environment and a make-to-order supply chain environment respectively. We show that an out-of-stock phenomenon that occurs earlier in a product's life cycle always leads to a significantly greater loss in total customer LV. We also demonstrate even a small lead time leads to a big loss in total customer LV.

Chapter 3 presents a model with heterogeneous population. We propose a four-segment model which considers the ex ante heterogeneity among customers in the tendency to be in tune with new developments and the tendency to influence (or be influenced by) others. Specifically, we segment customers into four types: type 1 customers are both innovators and global influencers, type 2 customers are both innovators and local influencers, type 3 customers are both imitators and global influencers, and type 4 customers are both imitators and local influencers. We characterize the closed-form expressions for adoption rate of each customer type. Based on them, we derive closed-form expressions for the customer PV, IV and LV as a function of product adoption time. We also investigate how PV, IV and LV vary with the adoption time and the innovation parameters. Finally, we analyze the impact of purchase acceleration on customer LV, and propose an algorithm based on the LV of marginal customers to optimally allocate free samples among customers.

This dissertation introduces the Binary Customer Satisfaction Model for addressing logistics issues. In typical logistics problems, the arrival of customers through a demand process is considered external to the management decisions. In practice, it is typically the case that customers will respond to changes is service policy by changing their behavior. The Binary Customer Satisfaction Model provides a simple customer behavior model that directly interacts with the service policy and provides for analysis of managerial insights.

The Binary Customer Satisfaction Model assigns customers to one of two satisfaction states. Satisfied customers have one demand rate, while unsatisfied customers have a different demand rate. The Binary Customer Satisfaction Model accommodates situations where satisfied customers demand more, as well as those when satisfied customers demand less, a possibility undertreated in existing literature. Satisfaction changes for each customer when the customer demands service. Satisfaction occurs when the customer receives service, while unsatisfaction occurs when the customer attempts to receive service but is unable to.

The Binary Customer Satisfaction Model is generally applicable to a wide range of logistics problems. In this dissertation, we consider the application of the model to the newsvendor inventory problem as well as an M/M/s queueing model without a buffer. We also briefly consider extensions to these models and how the Binary Customer Satisfaction Model can inform management of these extended cases.

Key to these insights is a well-defined concept of myopic management policies. This dissertation defines myopic policies in such a way to allow explicit comparison between optimal and myopic policies, and quantitatively present the value of considering the effect of service policy on future customer behavior.

In both the inventory and queueing contexts, we find that the Binary Customer Satisfaction Model gives two major insights. The first confirms the intuitive result that, if satisfied customers are more likely to arrive, then it is worthwhile for a manager to provide a level of service that would appear too high to a myopic manager, as future increases in customer demand will offset the additional cost. Similarly, if satisfied customers are less likely to arrive, the manager should prepare a lower level of service.

The second main insight is that it is not enough to simply observe the demand to find an optimal policy when customer behavior depends on service policies. Even with full knowledge of the demand, and beginning with the optimal demand level, a myopic manager will choose a suboptimal service policy, which will in turn create a suboptimal policy, which will cause the myopic manager to move further away from optimality. This spiraling effect makes clear the importance of not blindly making policies based on empirical observation of demand.