http://dspace.essec.fr/handle/essec/1 2026-10-25T09:13:04Z http://dspace.essec.fr/handle/essec/8305 Optimizing growth : Decision models for last-mile delivery and startup scaling Li, Yuanyuan Both large-scale e-commerce firms and small startups face unprecedented opportunities for growth, yet they must overcome significant operational challenges to scale efficiently. This dissertation focuses on two application domains—last-mile delivery operations and startup growth. It explores how operational decision models can be used to optimize growth strategies and guide resource allocation in fast-growing and competitive environments. Chapter 2 presents a structured review of emerging optimization problems for distribution in same-day delivery logistics. It first categorizes existing research, maps the delivery process, identifies sources of uncertainty, and reviews solution approaches. Then it compares solution structures obtained by applying the different formulations and discusses the corresponding implications. This chapter also provides managerial insights and outlines directions for future research. A paper based on this work has been accepted for publication in Computers and Operations Research. Chapter 3 studies the orienteering problem with stochastic and dynamic release dates in same-day delivery. Two reinforcement learning approaches based on value function approximation are proposed. One combines deterministic integer linear programs with a consensus function, while the other incorporates a two-stage stochastic program. The chapter shows that exact methods can be competitive in real-time routing environments. A paper based on this work is published in Transportation Science. The work was presented at the EURO 2022 in Helsinki,Finland, the POMS China 2023 in Zhejiang, China, and the TSL 2024 workshop in Nantes, France.Chapter 4 explores a delivery model that uses autonomous ground vehicles to transport parcels for direct deliveries from a central depot to a pickup station in each trip. This chapter analyzes the static and deterministic version of the problem and its variants, derives performance bounds, and evaluates the proposed heuristics through worstcase and numerical analyses. A paper based on this chapter is in progress.Chapter 5 investigates how startups can manage growth by balancing the costs of capacity adjustment with the need to reach a target market size. The analysis reveals that entrepreneurs should follow a strategic growth trajectory rather than pursuing aggressive expansion. In stable markets, increasing sales effort gradually is optimal,while in turbulent conditions, a more cautious and delayed ramp-up is preferred. A paper based on this chapter is in preparation. This work was presented at the POMS 2024 conference in Minneapolis, US, and the INFORMS Annual Meeting 2024 in Seattle, US. Chapter 6 concludes the dissertation by summarizing the major findings from each chapter and outlining opportunities for future research. 2025-01-01T00:00:00Z http://dspace.essec.fr/handle/essec/8304 Applications of Operations Research & Management in Product and Service Design Ramezanpour, Shaghayegh This thesis addresses decision-making problems in supply chain management and product line design under uncertainty, limited resources, and market complexity. The research develops optimizationbased approaches to improve operational resilience, supplier selection, and product design decisions while accounting for customer preferences, procurement policies, and economies of scale. Across three complementary studies, the thesis combines mathematical programming formulations with tailored exact solution methods and derives managerial insights through extensive computational experiments. Chapter 2 studies robust product fulfillment and return management for an omnichannel retailer facing supply chain disruptions that prevent customer access to stores, such as during a Covid-type pandemic. Specifically, the retailer operates one online channel and multiple physical stores, and disruptions result in temporary store closures as well as insufficient inventory in the online channel. To enhance operational resilience, the model allows selected physical stores to be reconfigured as online fulfillment stores, or DigiHubs, during the disruption. These DigiHubs function as decentralized stock depots that support online order fulfillment, process product returns, and help alleviate inventory shortages in the online channel. The key decision is to determine which physical stores should be reconfigured as DigiHubs during the disruption. Chapter 3 investigates supplier selection decisions in new product line design under customer choice behavior and supplier capacity limitations. Customers select their preferred product version according to a First-Choice model, while suppliers offer components with heterogeneous purchasing costs and production capacities, resulting in convex increasing procurement costs that capture supply scarcity conditions. The objective is to maximize firm profit by determining which product versions to include in the product line and by optimizing the trade-off between customer-driven revenue and component procurement costs. The problem is formulated as a mixed-integer linear program that becomes computationally challenging for large-scale instances. To address this issue, two tailored Benders decomposition reformulations are developed by alternatively assigning customer choice decisions to the master problem or the subproblem. Closed-form expressions for Benders cut coefficients are further derived to avoid repeatedly solving linear programming subproblems. Numerical experiments evaluate the computational performance of the proposed methods and provide managerial insights into supplier selection and product offering strategies. Chapter 4 focuses on feature selection in new product line design while accounting for economies of scale in procurement. The study determines feature levels, including pricing decisions, across multiple product versions to achieve an effective balance between customer satisfaction and production efficiency. Shared components across product versions generate purchasing discounts and economies of scale, introducing additional complexity into the design process. A mixed-integer linear programming formulation is proposed to maximize firm profit by managing the trade-off between product homogeneity, which reduces cost, and product heterogeneity, which better addresses customer preferences under a First-Choice part-worth utility model. Several families of valid inequalities are introduced to strengthen the formulation and improve computational efficiency. Extensive numerical analyses further explore the effects of purchasing discounts, market competitiveness, and customer heterogeneity on optimal design decisions, profitability, and the degree of product differentiation. Chapter 5 concludes the thesis by summarizing the main findings of the preceding chapters and briefly discussing potential research extensions and future directions. 2026-01-01T00:00:00Z http://dspace.essec.fr/handle/essec/8303 Three essays on trade exposure, innovation, and development finance Ma, Zipeng This dissertation contains three essays on the local effects of export demand and development finance. These effects vary with industrial structure, factor endowments, and human-capital allocation and may be obscured in national aggregates. Most “China shock” research treats Chinese exports as a shock to advanced-economy labor markets; these essays instead examine China’s own local economies and cities receiving Chinese development finance. The first essay studies this “Inverse China Shock.” Using Chinese firm-level data aggregated to cities and counties over 1998–2009, it applies fuzzy difference-in-differences and, separately, 2SLS using lagged ASEAN exports to the United States as an instrument for Chinese export exposure. The main IV estimates are positive for population-normalized manufacturing employment and aggregate manufacturing payroll. Employment effects are larger in low-skill and low-education groups, broadly consistent with Heckscher–Ohlin predictions, although payroll heterogeneity is less clear. The second essay studies regional innovation in Chinese manufacturing over 1998–2015. Greater export exposure significantly reduced city-level granted-patent intensity, a result reproduced by double/debiased machine-learning estimates; FDI remained positively associated with innovation, while the spatial estimates are consistent with positive neighboring-city spillovers. This pattern is consistent with a human-capital “Dutch disease” mechanism in which skilled labor may move away from innovation-intensive activity. The third essay compares geocoded Chinese and French aid projects in Morocco and Viet Nam over 2013–2020. At the displayed t+1 horizon following the first recorded change in project exposure, the main controlled staggered difference-in-differences estimates indicate an increase of about 0.010–0.016 exposure-index units in a constructed, revenue-normalized China-facing export-exposure index. The positive response is not sustained at t + 2 across specifications; French estimates are less precise and more estimator-sensitive, and the separate donor regressions do not establish a formal China–France difference. Across the three essays, external integration is associated with uneven local adjustments in manufacturing employment and payroll, regional innovation, and donor-facing export exposure. 2026-01-01T00:00:00Z http://dspace.essec.fr/handle/essec/8302 Stakeholder Scrutiny and Firms’ Strategic Actions on Environmental and Social Issues Singh, Shubham In my dissertation, I examine how external stakeholders exert pressure on firms to be socially responsible, and how firms respond to these pressures.Chapter 1. How does stakeholder scrutiny influence firms’ actions on environmental and social issues? Prior studies take an issue-centric approach whereby they focus on changes in firms’ actions on the specific social issue on which firms were scrutinized. I move beyond this issue-centric approach and highlight that scrutiny on a specific social issue may lead to an improvement in firms’ behavior on unrelated social issues. Firms may improve their performance on unrelated social issues to rebuild reputation and deflect the attention of wider stakeholders away from the issue on which firms were initially scrutinized. To find support for the above arguments, I exploit an exogenous shock that increased stakeholder scrutiny of firms in relation to the use of offshore tax havens. Using the difference-in-differences methodology, I analyze panel data on 256 firms from 2012 to 2020 and find that firms exposed to this shock increased their performance on unrelated social issues.Chapter 2. Socially responsible investment (SRI) funds claim to invest in firms with high social performance and engage with them to push them to improve their social performance. However, based on existing empirical evidence, it remains unclear whether they actually do so. Therefore, to assess whether SRI funds attempt to deliver on their promises, we examine how they respond when firms are targeted by activist hedge funds. As activist hedge funds often get firms to redirect their resources and managerial attention toward their priorities (which are often related to short-term financial performance), there is an impending risk that firms' social performance declines. Given this risk, it is important to ask whether SRI funds fight to push firms to prioritize social performance or they avoid investing in such firms to ensure that they have firms with high social performance in their portfolio. Using the difference-in-differences methodology, we analyze panel data on 462 firms from 2010 to 2022 and find evidence suggesting that SRI funds flee rather than fight. This finding illustrates that SRIs prioritize delivering on their promise of investing in socially responsible firms over their promise of engaging in activism to make firms socially responsible.Chapter 3. Does an increase in expectations of voluntary disclosure on environmental and social issues change firms' behavior on these issues? Prior studies suggest that such pressures are unlikely to be effective when firms can selectively increase disclosure on issues on which they perform well and avoid disclosure on issues on which they do not perform well. By engaging in selective disclosure, firms can give the impression that they care about transparency and simultaneously circumvent the need to improve their performance on environmental and social issues. Can self-regulation programs (which require/encourage firms to disclose on environmental and social issues) address this selective disclosure problem? Theoretically, it is unclear whether such programs can address this problem because there is a risk that when self-regulation programs introduce stringent reporting requirements to reduce selective disclosure, firms may exit such programs. Given this theoretical ambiguity, I abductively examine the impact of new disclosure requirements of the United Nations Global Compact (UNGC), a self-regulation program, on firms’ behavior on environmental and social issues. I employ a difference-in-differences design to analyze panel data on 646 firms and find that exposure to stringent disclosure requirements pushes firms to improve their performance on environmental and social issues. 2025-01-01T00:00:00Z