ZHANG Zhuoluo

Problem definition: About 30 million tons of plastic waste reach the oceans each year, most from low and middle-income coastal countries. We study novel business models of firms aiming to reduce ocean plastic pollution with a triple bottom line (TBL) objective–a weighted sum of profit, environmental and social impacts. These firms sell (a) plastic offsets and (b) segregated plastic. Methodology/results: We develop and analyze models where a firm partners with a local plastic recycling supply chain to sell (a) or (b) or both via collecting and recycling ocean-bound plastic. Considering additionality, i.e., that the firm can only sell plastic offsets based on recycled plastic that is additional to the plastic recycled without the firm’s presence, we solve the equilibrium outcomes by maximizing the firm’s TBL objective. For the special case of a for-profit firm, we show that additionality can decrease the firm’s social and environmental impacts when selling (a) only or when selling both (a) and (b). Additionality may also alter the effect of the local recycled plastic market (i.e., the number of collectors and the recycled plastic price) on the firm. We find similar insights under the TBL objective via a numerical study calibrated with real data. Managerial Implications: When firms decide whether to integrate and promote additionality, they must be careful since it may not only reduce their profit but also their social and environmental impacts. Moreover, we find that selling both (a) and (b) can generate a much higher TBL objective value than selling either one alone. We also find that firms employing a TBL objective can generate much larger environmental and social impacts with a slight reduction in profits than profit-maximizing firms. Our model and results provide insights into new initiatives for tackling ocean plastic pollution.

We consider a dynamic pricing problem for a consumer electronics trade-in program, where a firm acquires and re-sells multiple types of pre-owned (used) products over a finite selling horizon. The trade-in program offers two options: trade-in-for-cash, where customers sell their products to the firm and receive a cash payment, and trade-in-for-upgrade, where the customers exchange their products for new products at discounted prices. The firm sets trade-in prices (both cash rewards and new products' discounts) and resale prices to maximize its total expected profit. Customer arrivals follow independent Poisson processes and their choices on both used product trade-in and refurbished product purchase follow the Multinomial Logit (MNL) model. Given the challenge of solving the optimal policy using dynamic programming due to high dimensional state space, we develop simple and provably effective heuristic policies based on the solution to a deterministic upper-bound problem. We first propose a policy termed Static Control (SC) policy that computes prices before the start of the selling horizon. We show that its profit loss (relative to the optimal profit) is in the order of O(T^{1/2}) which matches that of the best possible stationary policy, where T is the number of selling periods. We then design a dynamic policy called the Batched-Adjustment Control (BAC) policy. Under BAC, the selling horizon is divided into different consecutive and disjoint batches for different products and the prices in one batch are updated based on the realized uncertainties in the previous batch. The profit loss of BAC is in the order of O(T^{1/3}). We numerically show that both policies perform well and BAC has superior performance over SC. Finally, we study three extensions of our model: Initial stocking of new products (for upgrade purposes), additional features of trade-in programs, and vertically differentiated products. We extend the dynamic policy and its theoretical performance analysis to all three extensions.

This paper investigates a retailer (he) provides various interventions to assist a supplier (she) with random production yield. The supplier faces constraints related to initial working capital and seeks external funding from a bank. We assume that the bank lacks precise information about the supplier’s production yield and operates with biased belief deviating from true values. We first build a benchmark model such that the retailer refrains from intervening, and find that information asymmetry may lead the bank to impose high interest rates on the supplier. This imposition elevates the supplier’s financial costs, subsequently diminishing the retailer’s profits. We then consider two other models: the first model, Retailer Credit Guarantee (RCG), entails the retailer committing to cover the supplier’s defaulted loans (termed credit guarantee), while the second model, Retailer Cost Sharing (RCS), involves the retailer sharing a portion of the supplier’s production costs. Our findings indicate that in scenarios where the bank underestimates the supplier’s production yield, either intervention can effectively reduce the supplier’s financial burdens, thereby potentially benefiting the retailer. Moreover, when the opportunity cost of the retailer is low and the bank underestimates the yield, the retailer tends to favor RCG over RCS with a sufficiently low cost-sharing proportion; conversely, this preference shifts when the opportunity cost of the retailer is great and the bank overestimates the yield. We also consider the scenario wherein the retailer simultaneously provides RCG and RCS to the supplier, and find that this combination strategy does not necessarily dominate RCG or RCS

We study initial stocking and inventory transshipment for a firm selling its product via multiple outlets. At the beginning of a selling season, the firm needs to distribute the product inventory to the outlets, which, in turn, supply customer demands. Customer demand to each store follows independent compound Poisson processes. Over the selling season, when needed, a re-distribution or transshipment of inventory can be carried out between the outlets. Hence, there are two decisions involved: the one-time stocking decision at the beginning of the season and the transshipment decision throughout the season. The firm aims to maximize the total expected profit over the season. We first employ stochastic dynamic programming (DP) to formulate the problem and characterize the optimal transshipment policy for a three-location problem. The optimal policy is quite complex and determined by a set of constants and functions (depending on the inventory status of outlets). In view of the challenge of using DP to solve the general multi-location problem, we design a simple heuristic policy that utilizes the solutions of easily solvable deterministic problems. By benchmarking against an upper bound problem, we show that the heuristic has a constant loss relative to the optimal profit, independent of the length of the selling season. We also develop an approximation with performance guarantee for computing the initial stocking levels used in the heuristic. Finally, a numerical study with model parameters calibrated by real data from a medical device company demonstrates the effectiveness of the heuristic.

We consider joint learning and pricing for an electronics trade-in platform which buys and sells multiple used electronics. The platform offers both trade-in-for-cash and trade-in-for-upgrade options and sets the corresponding prices to acquire the used products and their reselling prices after simple cleaning and refurbishing. Both supply and demand models are parametric. We propose two heuristics and analyze their regret bounds.

This study investigates the complexities of dynamic inventory management in the context of inventory-based financing (IBF), with a particular focus on inventory lines of credit (ILOC) in a multiperiod setting. Small and medium-sized enterprises (SMEs) often rely on IBF to overcome capital constraints, using their inventory as collateral to secure necessary funding. However, the fluctuating market value of this collateral introduces significant challenges for both inventory control and financing decisions. We develop a comprehensive dynamic inventory model that accounts for stochastic material prices and uncertain demand for the final product. The model seeks to maximize the firm’s expected total profit over a finite planning horizon by determining optimal inventory levels and borrowing amounts for both ILOC and short-term loans. Our analysis reveals that the optimal inventory policy is a state-dependent base-stock policy, which varies with the current material price and the firm’s equity level. Furthermore, we explore the conditions under which a firm may choose to liquidate inventory or seek additional short-term loans to manage the risk of price reductions that threaten the credit limit. Through extensive numerical studies, we examine the impact of key parameters such as the advance rate and interest rate on the firm’s optimal decisions and expected profitability. The findings offer valuable insights into the strategic interaction between operational and financial considerations in inventory management, particularly under conditions of capital constraints and price volatility.

Coming soon.