T he benefits of supplier diversification are well established for price-taking firms. In this paper, we investigate the benefits from supplier diversification for dual-sourcing duopolists. We consider a two-echelon supply chain in which suppliers sell components to buyers who produce and sell substitutable products. The suppliers' output processes are uncertain and modeled as having a proportional random yield. Buyers engage in a quantity-based Cournot competition. We find that an increase in supplier correlation leads to more correlated buyers' outputs and a decrease in their profits. In the presence of end-market competition, dual sourcing still brings value by reducing the inefficiency caused by random yield: Namely, when the suppliers' yield processes are strongly negatively correlated, dual sourcing increases the expected market output and improves the firms' profits over sole sourcing. However, unlike a monopolist firm, a duopolist does not necessarily allocate its supplier orders to minimize output variability. We generalize the main results to a two-stage order-quantityoutput-quantity game and to one with asymmetric suppliers.

S upply disruptions are all too common in supply chains. To mitigate delivery risk, buyers may either source from multiple suppliers or offer incentives to their preferred supplier to improve its process reliability. These incentives can be either direct (investment subsidy) or indirect (inflated order quantity). In this study, we present a series of models to highlight buyers' and suppliers' optimal parameter choices. Our base-case model has deterministic buyer demand and two possibilities for the supplier yield outcomes: all-or-nothing supply or partial disruption. For the all-or-nothing model, we show that the buyer prefers to only use the subsidy option, which obviates the need to inflate order quantity. However, in the partial disruption model, both incentives-subsidy and order inflation-may be used at the same time.Although single sourcing provides greater indirect incentive to the selected supplier because that avoids order splitting, we show that the buyer may prefer the diversification strategy under certain circumstances. We also quantify the amount by which the wholesale price needs to be discounted (if at all) to ensure that dual sourcing strategy dominates sole sourcing. Finally, we extend the model to the case of stochastic demand. Structural properties of ordering/subsidy decisions are derived for the all-or-nothing model, and in contrast to the deterministic demand case, we establish that the buyer may increase use of subsidy and order quantity at the same time.

Disruptions that temporarily interrupt production pose a significant risk for manufacturing firms. To manage this risk, firms can purchase interruption insurance and/or deploy operational measures such as storing inventory or taking preparedness actions that reduce the expected interruption length. In this study, we explore inventory, preparedness, and insurance in a two‐stage production chain that can experience disruptions at either the upstream or downstream stage. We analytically characterize an inventory‐only model and a preparedness‐only model in which the firm uses either inventory or preparedness effort to manage disruption risk, and a joint model in which the firm deploys both operational measures. We identify the relationships between the two operational measures within a stage and across the two stages. We also examine how insurance affects a firm's optimal deployment of and preference between the two operational measures. In addition to providing insights into the interaction of these three risk management measures, our results provide insights into the production chain design. For example, the firm can reduce its disruption risk management cost by allocating more production activity downstream (when possible) and this risk management benefit can, at times, outweigh the possible production cost increase associated with allocating more production downstream.

W e consider coordination issues in supply chains where supplier's production process is subject to random yield losses. For a simple supply chain with a single supplier and retailer facing deterministic demand, a pay back contract which has the retailer paying a discount price for the supplier's excess units can provide the right incentive for the supplier to increase his production size and achieve coordination. Building upon this result, we consider coordination issues for two other supply chains: one with competing retailers, the other with stochastic demand. When retailers compete for both demand and supply, they tend to over-order. We show that a combination of a pay back and revenue sharing mechanism can coordinate the supply chain, with the pay back mechanism correcting the supplier's under-producing problem and the revenue sharing mechanism correcting the retailers' over-ordering problem. When demand is stochastic, we consider a modified pay-back-revenue-sharing contract under which the retailer agrees to not only purchase the supplier's excess output (beyond the retailer's order), but also share with the supplier a portion of the revenue made from the sales of the excess output. We show that this contract, by giving the supplier additional incentives in the form of revenue share, can achieve coordination.

Patient no‐shows and late cancellations lead to clinic inefficiency, high clinic costs and low patient satisfaction. The two main strategies clinics employed to alleviate the adverse effects of no‐shows are overbooking and patient appointment reminders. Developing effective overbooking schedules depends on accurately predicting each patient’s no‐show probability, while developing effective reminder systems requires a patient‐level estimate of communication sensitivity. Current methods of estimating no‐show probabilities do not produce such patient‐level predictions. To remedy this, we develop a Bayesian nested logit model which utilizes appointment confirmation data and estimates individual‐level coefficients for patient‐specific predictors. A log‐likelihood comparison of model fit on 12 months of appointment data shows that the Bayesian model outperforms the standard logit model by about 30% improvement in model fit. Additionally, our Bayesian model allows categorization of patients based on their appointment confirmation behavior. Finally, using patient‐specific no‐show probabilities as an input to a simulated appointment scheduler we find that the Bayesian model improves clinic profit over the standard logit model. The benefit comes mainly from waiting cost reduction when no‐show probability is low and from physician overtime and idle time cost reduction when no‐show probability is high. Our study has two managerial implications. First, the Bayesian method allows customizing appointment reminder effort based on patient’s confirmation behavior. Second, the Bayesian method also allows improved overbooking scheduling especially in clinics that experience large patient throughput.