Tianhu Deng

Process Flexibility Design in Unbalanced Networks

This publication contains reprint articles for which IEEE does not hold copyright. Full text is not available on IEEE Xplore for these articles.

Optimal crude oil procurement under fluctuating price in an oil refinery

In this paper, we study the optimal procurement and operation of an oil refinery. The crude oil prices follow geometric Brownian motion processes with correlation. We build a multiperiod inventory problem where each period involves an operation problem such as separation or blending. The decisions are the amount of crude oils to purchase and the amount of oil products to produce. We employ approximate dynamic programming methods to solve this multiperiod multiproduct optimization problem. Numerical results reveal that this complex problem can be approximately solved with little loss of optimality. Further, we find that the approximate solution significantly outperforms a set of myopic policies that are currently used.

Statistical Learning of Service-Dependent Demand in a Multiperiod Newsvendor Setting

We study an inventory system wherein a customer may leave the sellers market after experiencing an inventory stockout. Traditionally, researchers and practitioners assume a single penalty cost to model this customer behavior of stockout aversion. Recently, a stream of researchers explicitly model this customer behavior and support the traditional penalty cost approach. We enrich this literature by studying the statistical learning of service-dependent demand. n nWe build and solve four models: a baseline model, where the seller can observe the demand distribution; a second model, where the seller cannot observe the demand distribution but statistically learns the demand distribution; a third model, where the seller can learn or pay to obtain the exact information of the demand distribution; and a fourth model, where demand in excess of available inventory is lost and unobserved. Interestingly, we find that all four models support the traditional penalty cost approach. This result confirms the use of a state-independent stockout penalty cost in the presence of demand learning. More strikingly, the first three models imply the same stockout penalty cost, which is larger than the stockout penalty cost implied by the last model.

Capacity allocation under downstream competition and bargaining

In this study, we consider a monopolistic supplier’s capacity-allocation problem under bargaining. The supplier can allocate one type of key element to either an external channel with a manufacturer, an internal channel, or both. The firms use the element to produce substitutable final products and compete in the product market. By building a stylized model, we characterize the equilibrium decisions under different channel choices. The conditions of the equilibrium channel choices are derived. We find that the supplier’s shared capacity increases with his bargaining power, but the manufacturer’s shared capacity decreases with her bargaining power. Meanwhile, the higher bargaining power may backfire on the manufacturer, because her loss from a decreased shared capacity may dominate her benefit from an increase in her bargaining power.

Data-Driven Pricing Strategy for Demand-Side Resource Aggregators

We consider a utility who seeks to coordinate the energy consumption of multiple demand-side flexible resource aggregators. For the purpose of privacy protection, the utility has no access to the detailed information of loads of resource aggregators. Instead, we assume that the utility can directly observe each aggregator’s aggregate energy consumption outcomes. Furthermore, the utility can leverage resource aggregator energy consumption via time-varying electricity price profiles. Based on inverse optimization technique, we propose an estimation method for the utility to infer the energy requirement information of aggregators. Subsequently, we design a data-driven pricing scheme to help the utility achieve system-level control objectives (e.g., minimizing peak demand) by combining hybrid particle swarm optimizer with mutation algorithm and an iterative algorithm. Case studies have demonstrated the effectiveness of the proposed approach against two benchmark pricing strategies—a flat-rate scheme and a time-of-use scheme.

Demand estimation under multi-store multi-product substitution in high density traditional retail

In large cities in emerging economies, traditional retail is present in a very high density, with multiple independently owned small stores in each city block. Consequently, when faced with a stockout, consumers may not only substitute with a different product in the same store, but also switch to a neighboring store. Suppliers may take advantage of this behavior by strategically supplying these stores in a coherent manner. We study this problem using consumer choice models. We build two consumer choice models for this consumer behavior. First, we build a Nested Logit model for the consumer choice process, where the consumer chooses the store at the first level and selects the product at the second level. Then, we consider an Exogenous Substitution model. In both models, a consumer may substitute at either the store level or the product level. Furthermore, we estimate the parameters of the two models using a Markov chain Monte Carlo algorithm in a Bayesian manner. We numerically find that the Nested Logit model outperforms the Exogenous Substitution model in estimating substitution probabilities. Further, the information on consumers’ purchase records helps improve the estimation accuracies of both the first-choice probabilities and the substitution probabilities when the beginning inventory level is low. Finally, we show that explicitly including such substitution behavior in the inventory optimization process can significantly increase the expected profit.

Process flexibility design in unbalanced networks

Several design guidelines and flexibility indices have been developed in the literature to inform the design of flexible production networks. In this paper, we propose additional flexibility design guidelines for unbalanced networks, where the numbers of plants and products are not equal, by refining the well-known Chaining Guidelines. We study symmetric networks, where all plants have the same capacity and product demands are independent and identically distributed, and focus mainly on the case where each product is built at two plants. We also briefly discuss cases where (1) each product is built at three plants and (2) some products are built at only one plant. An extensive computational study suggests that our refinements work very well for finding flexible configurations with minimum shortfall in unbalanced networks.