Tinglong Dai

Agent Reasoning in Negotiation

Negotiation has been studied in different communities both scientific and communities of practice. The social sciences (see chapters by Martinovski, Albin and Druckman, Koeszegi and Vetschera, this volume for example) and the mathematical sciences have investigated different aspects of negotiation with different goals: the goals of the social sciences are to understand the factors and reasoning processes that underlie human negotiation behavior. The goal of the mathematical sciences is to formulate mathematical models that capture elements of negotiation. Further, the mathematical models can be divided into analytic models (economic, operations research etc) and computational models. The aim of the analytic models is to provide guarantees of their behavior, characterizations of optimality, or provide managerial guidance to optimize negotiation activity. The computational models aim to provide computational tractability through approximation algorithms and heuristics. Most crucially, the computational research aims to have the models implemented in autonomous processes, called agents, that are able to incorporate realistic factors of negotiation (e.g. argumentation, information seeking, and cognitive factors) and engage in negotiations in a decentralized manner. Such agent models promise to contribute to our understanding of human information processing in negotiation. Additionally, they could be used for decision support of human decision makers. In the long run, such models can even become substitutes for human negotiators. In this chapter we will provide a selective review of the most important works in the analytic and computational negotiation literature, point out some differences and synergies and provide pointers to open questions and future research.

An acquisition policy for a multi-supplier system with a finite-time horizon

We study the problem of a manufacturer who outsources a single product to multiple capacitated suppliers. Given a pool of potential suppliers, a decision has to be made about which supplier(s) to use, and how much and how often to purchase from each selected supplier. We build an EOQ-type model with a finite-time horizon, which differs from the existing model in the literature with an infinite-time horizon. We design a dynamic programming and a scatter search algorithm to solve the problem, and provide computational results to show the effectiveness of the proposed methods.

Service level differentiation in multi-robots control

In this paper we explore the effects of service level differentiation on a multi-robot control system. We examine the premise that although long interaction time between robots and operators hurts the efficiency of the system, as it generates longer waiting time for robots, it provides robots with longer neglect time and better performance benefiting the system. In the paper we address the problem of how to choose the optimal service level for an operator in a system through a service level differentiation model. Experimental results comparing system performance for different values of system parameters show that a mixed strategy is a general way to get optimal system performance for a large variety of system parameter settings and in all cases is no worse than a pure strategy.

Contracting for On-Time Delivery in the U.S. Influenza Vaccine Supply Chain

Although influenza vaccine shortage is often attributed to low supply, it has been observed that even with abundant supply, a major shortage can still occur because of late delivery. In this paper, motivated by the influenza vaccine industry, we study a supply chain contracting problem in the presence of uncertainties surrounding design, delivery, and demand of the influenza vaccine. In this supply chain, a manufacturer has insufficient incentive to initiate at-risk early production prior to the design freeze because it is a retailer who reaps the most benefits from selling more vaccines delivered on time. Anticipating that late delivery will lead to potential loss in demand, the retailer tends to reduce the order size, which further discourages the manufacturer from making an effort to improve its delivery performance. To break this negative feedback loop, a supply contract needs to achieve two objectives: incentivize at-risk early production and eliminate double marginalization. We find that two commonly observed supply contracts in practice, the delivery-time-dependent quantity flexibility (D-QF) contract and the late-rebate (LR) contract, may fail to coordinate the supply chain because of the tension between these two objectives. To resolve such a tension, we construct a buyback-and-late-rebate (BLR) contract and show that a properly designed BLR contract can not only coordinate the supply chain but also can provide full flexibility of profit division between members of the supply chain. Numerical experiments further demonstrate that the BLR contract significantly improves supply chain efficiency compared to the contracts used in the industry.

Modeling Power Distance and Individualism/Collectivism in Negotiation Team Dynamics

It has been documented in the social sciences that cultural factors affect how people negotiate and behave in negotiations. Despite the importance of culture in the business world and politics, there is a lack of computational models that help to analyze how cultural factors affect negotiation. Moreover, while many negotiations take place between teams, there is a dearth of computational models for team negotiations. In this paper we present the first attempt to provide a computational model which takes into account cultural factors in a team negotiation setting. The model considers how two important cultural dimensions, power distance and individualism/collectivism, affect team negotiation dynamics and negotiation outcomes. We conducted experiments in high/low intra-team conflict scenarios. The results are compatible with social sciences findings from team decision making.

A game theoretic queueing approach to self-assessment in human-robot interaction systems

This paper presents a queueing model that addresses robot self-assessment in human-robot-interaction systems. We build the model based on a game theoretic queueing approach, and analyze four issues: 1) individual differences in operator skills/capabilities, 2) differences in difficulty of presenting tasks, 3) trade-off between human interaction and performance and 4) the impact of task heterogeneity in the optimal service decision-making and system performance. The subsequent analytical and numerical exploration helps understand the way the decentralized decision-making scheme is affected by various service environments.

Equity-Based Incentives and Supply Chain Buy-Back Contracts*

We analyse the effect of equity-based incentives in a supply chain with a downstream firm and an upstream supplier. By using the operational decision as a signal to influence external investors’ beliefs, the downstream firm’s manager intends to maximize a convex combination of the interim share price and the terminal cash flows. We show that equity-based incentives create a side-effect. Specifically, with a universal buy-back contract, the deadweight loss of signalling induced by equity-based incentives could spread throughout the supply chain and cause chain-wide damages. To mitigate such undesirable consequences, we propose a new mechanism to eliminate the inefficiency. We derive the optimal mechanism that maximizes the downstream firm’s profits subject to the constraint that the supply chain efficiency is not undermined. In contrast to the full-information benchmark, this mechanism gives positive surplus to the supplier.

Automated Bilateral Multiple-issue Negotiation with No Information About Opponent

In this paper, we investigate offer generation methods for automated negotiation on multiple issues with no information about the opponents utility function. In existing negotiation literature, it is usually assumed that an agent has full information or probabilistic beliefs about the other agents utility function. However, it is usually not possible for agents to have complete information about the other agents preference or accurate probability distributions. We prove that using an alternating projection strategy, it is possible to reach an agreement in general automated multi-attribute negotiation, where the agents have nonlinear utility functions and no information about the utility function of the other agent. We also prove that rational agents do not have any incentive to deviate from the proposed strategy. We further present simulation results to demonstrate that the solution obtained from our protocol is quite close to the Nash bargaining solution.

Toward a Unified Negotiation Framework: Leveraging Strengths in Behavioral and Computational Communities

While there has been a large body of negotiation literature in both Behavioral Science (behavioral) and Artificial Intelligence/Game Theory (computational) communities, there has not been an attempt to bridge the two communities to our best knowledge. In this chapter, we compare and contrast the characteristics of behavioral and computational literature in negotiation. We propose that incorporating the strengths of two types of literature are valuable in expanding the horizon of research outlook.

Technical Note—Impact of Inventory on Quota-Bonus Contracts with Rent Sharing

We study the impact of limited inventory on optimal sales-force compensation contracts. We adopt a principal-agent framework, characterized by limited liability and rent sharing with the agent. A commonly invoked assumption in the inventory management literature is that the demand distribution satisfies the increasing failure rate (IFR) property. Under this assumption, however, past research has established that a quota-bonus contract—a widely adopted sales-force compensation contract in the practice—cannot sustain in equilibrium. We show that because of demand censoring in the presence of limited inventory (i.e., demand realizations higher than the inventory level are unobservable), a quota-bonus contract is the optimal equilibrium contract, and it exists, even for a demand distribution with the IFR property. Since most well-known distributions satisfy the IFR property, and inventory constraints are operative in many real-world situations, our results significantly extend the scope of the optimality of quota-bonus contracts and underscore the importance of considering the inventory aspect while making sales-force compensation decisions.