Rebalancing Autonomous Vehicles using Deep Reinforcement Learning

The shared autonomous mobility-on-demand (AMoD) system is a promising business model in the coming future which provides a more efficient and affordable urban travel mode. However, to maintain the efficient operation of AMoD and address the demand and supply mismatching, a good rebalancing strategy is required. This paper proposes a reinforcement learning-based rebalancing strategy to minimize passengers’ waiting in a shared AMoD system. The state is defined as the nearby supply and demand information of a vehicle. The action is defined as moving to a nearby area with eight different directions or staying idle. A 4.6 4.4 km2 region in Cambridge, Massachusetts, is used as the case study. We trained and tested the rebalancing strategy in two different demand patterns: random and first-mile. Results show the proposed method can reduce passenger’s waiting time by 7% for random demand patterns and 10% for first-mile demand patterns.

Impact of quality decisions on information sharing with supplier encroachment

This study examines a retailer's decision to share market demand information in a supply chain wherein a supplier sells a product with a certain level of quality to a retailer, who then resells it to the end consumer. It also considers whether a supplier should establish a direct selling channel by incurring a fixed entry cost to compete with the retail channel. Although conventional wisdom indicates that a retailer may voluntarily disclose information under ex-ante supplier encroachment, our results show how and why a retailer may share information with the supplier who encroaches on the retail market with a decision on quality. Specifically, our findings reveal that information sharing is beneficial to the retailer when the quality cost coefficient is low and entry cost is relatively low, even under encroachment by the supplier. Moreover, the retailer may prefer to disclose demand information to the supplier if the quality cost coefficient is low, even when the entry cost is high, under non-encroachment. Interestingly, we found that the supplier prefers to encroach if the retailer shares demand information when the entry cost is moderate. Further, we found that the retailer has a higher incentive to share information under supplier encroachment compared with non-encroachment. These results are in sharp contrast with the extant literature.

Simulation Research on Information-Sharing Value of Two-Level Supply Chain

In the two-stage supply chain model, the incentive effect to the supplier’s sharing of demand information and performance evaluation and the effect of various parameters on the incentive effect of the supply chain are studied through a multiagent simulation model constructed for the purpose. It is found that the incentive coefficient of demand information-sharing degree, the number of selected suppliers, the order allocation coefficient, and the order proportion are positively related to the incentive effect of demand information sharing. So, the greater the demand information sharing is, the greater the impact of these parameters on the incentive effect is. Based on the demand information sharing, the supplier performance evaluation rules are shared, and when the actual evaluation rules are inconsistent with the supplier’s expectations, the incentive effect is further enhanced. Other parameters do not affect the incentive effect of demand information sharing and performance evaluation rule sharing.

Dynamic Pricing and Inventory Control with Fixed Ordering Cost and Incomplete Demand Information

We consider the periodic review dynamic pricing and inventory control problem with fixed ordering cost. Demand is random and price dependent, and unsatisfied demand is backlogged. With complete demand information, the celebrated [Formula: see text] policy is proved to be optimal, where s and S are the reorder point and order-up-to level for ordering strategy, and [Formula: see text], a function of on-hand inventory level, characterizes the pricing strategy. In this paper, we consider incomplete demand information and develop online learning algorithms whose average profit approaches that of the optimal [Formula: see text] with a tight [Formula: see text] regret rate. A number of salient features differentiate our work from the existing online learning researches in the operations management (OM) literature. First, computing the optimal [Formula: see text] policy requires solving a dynamic programming (DP) over multiple periods involving unknown quantities, which is different from the majority of learning problems in OM that only require solving single-period optimization questions. It is hence challenging to establish stability results through DP recursions, which we accomplish by proving uniform convergence of the profit-to-go function. The necessity of analyzing action-dependent state transition over multiple periods resembles the reinforcement learning question, considerably more difficult than existing bandit learning algorithms. Second, the pricing function [Formula: see text] is of infinite dimension, and approaching it is much more challenging than approaching a finite number of parameters as seen in existing researches. The demand-price relationship is estimated based on upper confidence bound, but the confidence interval cannot be explicitly calculated due to the complexity of the DP recursion. Finally, because of the multiperiod nature of [Formula: see text] policies the actual distribution of the randomness in demand plays an important role in determining the optimal pricing strategy [Formula: see text], which is unknown to the learner a priori. In this paper, the demand randomness is approximated by an empirical distribution constructed using dependent samples, and a novel Wasserstein metric-based argument is employed to prove convergence of the empirical distribution. This paper was accepted by J. George Shanthikumar, big data analytics.

Sharing Manufacturer’s Demand Information in a Supply Chain with Price and Service Effort Competition

Problem definition: This paper investigates the issue of sharing the private demand information of a manufacturer that sells a product to retailers competing on prices and service efforts. Academic/practical relevance: In the existing literature, which ignores service effort competition, it is known that demand signaling induces an informed manufacturer to distort the wholesale price downward, which benefits the retailers, and so, they do not have any incentive to receive the manufacturer’s private information. In practice, many manufacturers share demand information with their retailers that compete on prices and service efforts (e.g., demand-enhancing retail activities), a setting that has not received much attention from the literature. Methodology: We develop a game-theoretic model with one manufacturer selling to two competing retailers and solve for the equilibrium of the game. Results: We show how an informed manufacturer may distort the wholesale price upward or downward to signal demand information to the retailers, depending on the cost of service effort, the intensity of effort competition, and the number of uninformed retailers. We fully characterize the impact of such wholesale price distortion on the firms’ incentive to share information and derive the conditions under which the manufacturer shares information with none, one, or both of the retailers. We derive conditions under which a higher cost of service effort makes the retailers or the manufacturer better off. Managerial implications: Our results provide novel insights about how service effort competition impacts the incentives for firms in a supply chain to share a manufacturer’s private demand information. For instance, when the cost of effort is high or service effort competition is intense, a manufacturer should share information with none or some, but not all, of the retailers.

Optimal Ordering Policy for Retailers with Bayesian Information Updating in a Presale System

In this study, we investigate inventory allocation and pricing strategies for retailers by incorporating demand information into the issue of inventory allocation during the presale period. In a presale system, retailers offer presale goods at a price lower than the retail price. By offering products at a discount, retailers may attract additional demand. In addition, this system enables retailers to reduce the uncertainty of market demand and establish a strategy for inventory allocation based on the results of presales. A Bayesian approach was employed to analyze and update demand information, and inventory allocation was formulated as a newsvendor problem to determine the optimal policy that maximizes retailer profit . A numerical analysis was conducted to validate the effectiveness of the proposed strategy. Results suggest that the proposed strategies can support retailers by more accurately predicting demand and achieving higher profits with less inventory. Furthermore, retailers can experience greater benefits from risk-averse customers than from risk-neutral customers.

Pricing strategies of low-carbon enterprises in the Yellow River Basin considering demand information and traceability services

PurposeThe purpose of this paper is to solve new pricing issues faced by low-carbon companies in the Yellow River Basin, which is caused by the change of key pricing factors in the mixed appliance background of Big Data and blockchain, such as product quality and carbon-emission reduction CER level (hereafter, CER level).Design/methodology/approachWe choose a low-carbon supply chain with a low-carbon manufacturer and a retailer as our research object. Then, we propose that using the ineffective effect of the CER level and the quality and safety level to reflect the relationships among the CER level, the quality and safety level and the market demand is more suitable in the new environment. Based on these, we revise the demand equation. Afterwards, by using Stackelberg game, four cost-sharing situations and their pricing rules are analyzed.FindingsResults indicated that in the four cost-sharing situations, the change trends and the magnitudes of the best retail prices were not affected by the changes of the inputs of the demand information and the traceability services costs (hereafter, DITS costs), the proportion about retailer's DITS costs undertaken by the manufacturer, the ineffective effect coefficient of the CER level and the quality and safety level and the cost optimization coefficient. However, the cost-sharing situations could affect the change magnitudes of the best revenues.Originality/valueThis paper has two main contributions. First, this paper proposes a demand function that is more suitable for the mixed appliance background of Big Data and blockchain. Secondly, this paper improves the cost-sharing model and finds that demand information sharing and traceability service sharing have different impacts on key pricing factors of low-carbon product. In addition, this research provides a theoretical reference for low-carbon supply chain members to formulate pricing strategies in the new background.