With the emergence of the Internet as a global structure for communication and interaction, many “business to consumer” and “business to business” applications have migrated online, thus increasing the need for software agents that can act on behalf of people, institutions or companies with private and often conflicting interests. The design of such agents, and the protocols (i.e., mechanisms) through which they interact, has therefore naturally become an important research theme. Classical mechanism design techniques from the economics literature do not account for the costs entailed with the full revelation of preferences that they require. The aim of this thesis is to investigate how to design mechanisms that only require the revelation of partial preference information and are applicable in any mechanism design context. We call this partial revelation mechanism design. Reducing revelation costs is thus our main concern. With only partial revelation, the designer has some remaining uncertainty over the agents’ types, even after the mechanism has been executed. Thus, in general, the outcome chosen will not be optimal with respect to the designer’s objective function. This alone raises interesting questions about which (part of the) information should be elicited in order to minimize the degree of sub-optimality incurred by the mechanism. But this sub-optimality of the mechanism’s outcome choice function has additional important consequences: most of the results in classical mechanism design which guarantee that agents will reveal their type truthfully to the mechanism rely on the fact that the optimal outcome is chosen. We must therefore also investigate if, and how, appropriate incentives can be maintained in partial revelation mechanisms. We start by presenting our own model for partial revelation mechanism design. Our second contribution is a negative one regarding the quasi-impossibility of implementing partial revelation mechanisms with exact incentive properties. The rest of the thesis shows, in different settings, how this negative result can be bypassed in various settings, depending on the designer's objective (e.g., social welfare, revenue...) and the interaction type (sequential or one shot). Finally, we study how the approximation of the incentive properties can be further improved when necessary, and in the process, introduce and proves the existence of a new equilibrium concept.

With the emergence of the Internet as a global structure for communication and interaction, many “business to consumer” and “business to business” applications have migrated online, thus increasing the need for software agents that can act on behalf of people, institutions or companies with private and often conflicting interests. The design of such agents, and the protocols (i.e., mechanisms) through which they interact, has therefore naturally become an important research theme. Classical mechanism design techniques from the economics literature do not account for the costs entailed with the full revelation of preferences that they require. The aim of this thesis is to investigate how to design mechanisms that only require the revelation of partial preference information and are applicable in any mechanism design context. We call this partial revelation mechanism design. Reducing revelation costs is thus our main concern. With only partial revelation, the designer has some remaining uncertainty over the agents’ types, even after the mechanism has been executed. Thus, in general, the outcome chosen will not be optimal with respect to the designer’s objective function. This alone raises interesting questions about which (part of the) information should be elicited in order to minimize the degree of sub-optimality incurred by the mechanism. But this sub-optimality of the mechanism’s outcome choice function has additional important consequences: most of the results in classical mechanism design which guarantee that agents will reveal their type truthfully to the mechanism rely on the fact that the optimal outcome is chosen. We must therefore also investigate if, and how, appropriate incentives can be maintained in partial revelation mechanisms. We start by presenting our own model for partial revelation mechanism design. Our second contribution is a negative one regarding the quasi-impossibility of implementing partial revelation mechanisms with exact incentive properties. The rest of the thesis shows, in different settings, how this negative result can be bypassed in various settings, depending on the designer's objective (e.g., social welfare, revenue...) and the interaction type (sequential or one shot). Finally, we study how the approximation of the incentive properties can be further improved when necessary, and in the process, introduce and proves the existence of a new equilibrium concept.

My dissertation studies the optimal design of institutions and information structures for different objectives of a designer or a social planner. The questions addressed are interesting both from a theoretical point of view, and in terms of their real-life applications. The first chapter of the dissertation focuses on supermodular mechanism design in environments with arbitrary finite type spaces and interdependent valuations. In these environments, the designer may have to use Bayesian equilibrium as a solution concept, because ex post implementation may not be possible. We propose direct Bayesian mechanisms that are robust to certain forms of bounded rationality while controlling for equilibrium multiplicity. In quasi-linear environments with informational and allocative externalities, we show that any Bayesian mechanism that implements a social choice function can be converted into a supermodular mechanism that also implements the original decision rule. The proposed supermodular mechanism can be chosen in a way that minimizes the size of the equilibrium set, and we provide two sets of sufficient conditions to this effect: for general decision rules and for decision rules that satisfy a certain requirement. This is followed by conditions for supermodular implementation in unique equilibrium. The second chapter looks at the incentives of a revenue-maximizing seller (designer) who discloses information to a number of interacting bidders (agents). In particular, the designer chooses the level of precision with which agents can infer the quality of a common-value object from their privately observed signals. We restrict attention to the second-price sealed-bid auction format. If the seller has perfect commitment power and can choose the precision level before observing the quality of the object, in the presence of any small cost to precision it is ex ante optimal for her to choose completely uninformative signals. For the case when the seller chooses the precision level after observing the quality of the object, we characterize pooling, partial pooling, and separating equilibria. We show that in this setting the cost associated with precision can be viewed as a form of commitment device: if costs are too low, the best pooling equilibrium ceases to exist as the high type seller is too tempted to separate. Thus, the seller ends up with a lower ex ante expected payoff than in the case when cost parameters are above a certain threshold. The third chapter of this dissertation studies the optimal choice of information structure from the perspective of a designer maximizing a certain objective function. Generally speaking, there are two ways of creating incentives for interacting agents to behave in a desired way. One is by providing appropriate payoff incentives, which is the subject of mechanism design. The other is by choosing the information that agents observe, which we refer to as information design. We consider a model of symmetric information where a designer chooses and announces the information structure about a payoff relevant state. The interacting agents observe the signal realizations, update their beliefs, and take actions which affect the welfare of both the designer and the agents. We characterize the general finite approach to deriving the optimal information structure --- the one that maximizes the designer's ex ante expected utility subject to agents playing a Bayes Nash equilibrium. We then apply the general approach to a symmetric two state, two agent, and two actions environment in a parameterized underlying game and fully characterize the optimal information structure. It is never strictly optimal for the designer to use conditionally independent private signals. The optimal information structure may be a public signal, or may consist of correlated private signals. Finally, we examine how changes in the underlying game affect the designer's maximum payoff. This exercise provides a joint mechanism/information design perspective. / text

University of Macau / Faculty of Science and Technology / Department of Electromechanical Engineering

<p>I characterize optimal long term contracts offered by a monopolist to a buyer whose private valuation evolves according to a branching process with privately known transition probability. The optimal contract can be implemented in a simple way, and presents the buyer with a tradeoff between a high initial fixed fee and low future prices. In an interaction with a long time horizon, the relationship will terminate prematurely with probability close to one. Optimal mechanisms are quite different from models in which the transition probability is known, and the buyer's private information is his initial valuation. Optimal contracts resemble the structure of term life insurance contracts, and have features similar to actual interactions between retailers and suppliers.</p> / Dissertation

Doctor of Philosophy (Economics) / Auctions are an ancient economic institution. Since Vickrey (1961), the development of auction theory has lead to an extremely detailed description of the often desirable characteristics of these simple selling procedures, in the process explaining their enduring popularity. Given the pervasiveness of auctions, the question of how a seller should engineer the rules of these mechanisms to maximize her own profits is a central issue in the organization of markets. The seminal paper of Myerson (1981) shows that when facing buyers with Independent Private Values (IPVs) a standard auction with a specifically selected reserve price (or prices) is optimal, that is, maximizes a seller's expected profits among all conceivable selling mechanisms. In this model, it is assumed that the buyers have perfect information as to the existence of gains from trade. We shall argue that the consequences of this assumption for the design of the optimal auction are not well understood, which motivates our analysis. The three essays of this thesis relax the `known seller valuation' assumption by examining the optimal auction program when the seller (and principal) holds private information representing her reservation value for the good. In the first essay we provide an original technique for comparing ex ante expected profits across mechanisms for a seller facing N>1 potential buyers when all traders hold private information. Our technique addresses mechanisms that cannot be ranked point-by-point through their allocation rules using the Revenue Equivalence Theorem. We find conditions such that the seller's expected profits increase in the slope of each buyer's allocation probability function. This provides new intuition for the fact that a principal does not benefit from holding private information under risk neutrality. Monopoly pricing induces steep probability functions so the seller/principal benefits from announcing a fixed price, and implicitly her private information. An application is presented for the well known k double auction of the bilateral trade literature. In the second and third essays of this thesis, we extend the above framework to allow for informational externalities. Specifically, we allow for the situation in which the seller's private information represents a common value component in buyers' valuations. Thus the seller's private information (say regarding the quality of the good) is of interest to bidders independently of any strategic effects. In recent work Cai, Riley and Ye (2007) have demonstrated that a seller who holds private information about the quality of a good faces an extra consideration in designing an auction; the reserve price signals information to bidders. In a separating equilibrium signalling is costly in the sense that reserves are higher than would be optimal under complete information. We examine the returns to the seller in an English auction from using different types of secret reserve regimes. We find that immediate disclosure of a reserve is preferable to announcement after the auction in the form of a take-it-or-leave-it offer to the winning bidder. Sale occurs less often during the auction for a given reserve price strategy under secret reserve regimes, which increases the incentive for the seller to report more favourable information though the reserve price offer. Separating equilibria involving later announcement therefore generate even lower expected profits to the seller (signalling is more costly) than under immediate disclosure. In the third essay we compare the benchmark signalling equilibrium of immediate disclosure to a screening regime which we call the Right of Refusal. In this extreme form of a secret reserve the seller never announces the reserve price, she simply accepts or rejects the auction price. We find that the Right of Refusal dominates immediate disclosure if the seller's valuation is a sufficient statistic for the private information of interest. Thus a seller with market-relevant private preference information can benefit from not exercising monopoly price setting power. The result also provides conditions under which a competitive screening equilibrium is more efficient than a signalling mechanism. Broadly speaking, screening is better when the common value aspect in the preferences of the informed and uninformed parties are `aligned', and potential gains from trade to the uninformed party are significant. We believe this conclusion to be of particular interest to the design of privatization schemes.

The MEMS technology has been developed for many years. Micro-assembly becomes more and more important in this years. Microfasteners play a critical role in micro-assembly due to their simple configuration and easy to be fabricated. The objective of this study is to utilize the phenomenon observed, that fits the definition of the degree of freedom of general mechanisms, and to synthesize new microfasteners employing the concept of mechanism design. Firstly, microfasteners are analyzed based on the need. After synthesizing new microfasteners in micro-scale, they are transformed into compliant mechanisms in micro-scale. By establishing geometry, FEM is applied to simulate the situation of insertion and retention, and to check if bucking will happen. Finally, the retention force is improved by changing the width of links. The insertion force, retention force, and stress are considered for selecting feasible configurations. An example of microgear set mechanism is introduced to demonstrate the assembly of the microgear set after microfabrication process with large backlash.

Soft automation employs multi-functional robotic manipulators where dexterity, versatility and reconfigurability are now becoming key issues. For manipulators used in these circumstances, hyper-redundancy in configuration is essential. A typical application envisaged, is manipulator movement in unpredictable surroundings such as navigation amongst branches and foliage in automated fruit picking. This requires motion of a manipulator in an unmapped dynamic environment to find its way to the goal i.e. the fruit. To date, centralized control architectures have proven impractical for the control of redundant systems because the full environment needs to be mapped, there are time constraints on the computation required and, simultaneously, algorithms for obstacle avoidance must be enacted. Decentralized control of manipulators complements modular construction because the need for a full environment reconstruction in one master controller is bypassed by having localized sub-goals for each module. Time constraints are removed because the control algorithms are much simpler. Obstacle avoidance is localized. Manipulators constructed modularly are effective because they allow for reconfiguration and ease of fault diagnosis. For modular manipulators to be a more effective option as a subclass of robots, the conditions under which the interactive movements of modules are stable become a major issue. When a general review of hyper-redundant manipulators was undertaken, no published implementation of Modular Decentralized Control (MDC) was discovered. This thesis explored the use of a modular decentralized technique to create stable control of a redundant manipulator system. The computational burden was minimized by restricting inverse kinematics to within each module. Advantages of the approach taken were the ease of implementation of obstacle avoidance, reconfigurability and fault tolerance. Having firstly simulated a MATLAB version of stable motion using MDC on a modular manipulator with up to six identical modules, the technique was extended with state space analysis to redefine the limits of stable control of a hyper-redundant manipulator. The MDC study mapped motion profile types that were dependant on the initial manipulator configuration and goal position and, thereby, investigated possible instabilities in the system. A two-link, single degree of freedom system was initially explicated followed by an extension of the stability analysis to an n-module two degree of freedom system. A stability theory utilizing decentralized control was formed. Simulation results showed dynamic motion, path generation and obstacle avoidance capabilities in unmapped environments to be stable. The modeling redefined the bounds of stable control, showing that classical stability via Root Locus, now required only two roots from the characteristic equation to be stable for a selection of path trajectory to the goal to be found. The remaining roots could be unstable in traversing to the goal and settling at a marginal stability point when the goal was reached. The marginal stability was a reflection of the pseudo-independence given to each module in seeking the goal and differed radically from a standard Root Locus analysis and interpretation of stability. A hyper-redundant Reconfigurable Modular Manipulator System (RMMS) was designed and built to implement the MDC technique in a real world environment. From an initial design, five modules were constructed and control algorithms embedded appropriate to their position in a five-segment robotic manipulator. A stereoscopic vision system was attached to the end of the manipulator which supplied real time data on a goal in 3D Cartesian space. The data was supplied to the first module of the arm and subsequently to all others by localized homogenous transformation. The manipulator was tested for goal seeking, path following, obstacle avoidance, fault tolerance and reconfigurability. The arm produced stable motion and satisfied the criteria as hypothesized in the theory.

Doctor of Philosophy (Economics) / Auctions are an ancient economic institution. Since Vickrey (1961), the development of auction theory has lead to an extremely detailed description of the often desirable characteristics of these simple selling procedures, in the process explaining their enduring popularity. Given the pervasiveness of auctions, the question of how a seller should engineer the rules of these mechanisms to maximize her own profits is a central issue in the organization of markets. The seminal paper of Myerson (1981) shows that when facing buyers with Independent Private Values (IPVs) a standard auction with a specifically selected reserve price (or prices) is optimal, that is, maximizes a seller's expected profits among all conceivable selling mechanisms. In this model, it is assumed that the buyers have perfect information as to the existence of gains from trade. We shall argue that the consequences of this assumption for the design of the optimal auction are not well understood, which motivates our analysis. The three essays of this thesis relax the `known seller valuation' assumption by examining the optimal auction program when the seller (and principal) holds private information representing her reservation value for the good. In the first essay we provide an original technique for comparing ex ante expected profits across mechanisms for a seller facing N>1 potential buyers when all traders hold private information. Our technique addresses mechanisms that cannot be ranked point-by-point through their allocation rules using the Revenue Equivalence Theorem. We find conditions such that the seller's expected profits increase in the slope of each buyer's allocation probability function. This provides new intuition for the fact that a principal does not benefit from holding private information under risk neutrality. Monopoly pricing induces steep probability functions so the seller/principal benefits from announcing a fixed price, and implicitly her private information. An application is presented for the well known k double auction of the bilateral trade literature. In the second and third essays of this thesis, we extend the above framework to allow for informational externalities. Specifically, we allow for the situation in which the seller's private information represents a common value component in buyers' valuations. Thus the seller's private information (say regarding the quality of the good) is of interest to bidders independently of any strategic effects. In recent work Cai, Riley and Ye (2007) have demonstrated that a seller who holds private information about the quality of a good faces an extra consideration in designing an auction; the reserve price signals information to bidders. In a separating equilibrium signalling is costly in the sense that reserves are higher than would be optimal under complete information. We examine the returns to the seller in an English auction from using different types of secret reserve regimes. We find that immediate disclosure of a reserve is preferable to announcement after the auction in the form of a take-it-or-leave-it offer to the winning bidder. Sale occurs less often during the auction for a given reserve price strategy under secret reserve regimes, which increases the incentive for the seller to report more favourable information though the reserve price offer. Separating equilibria involving later announcement therefore generate even lower expected profits to the seller (signalling is more costly) than under immediate disclosure. In the third essay we compare the benchmark signalling equilibrium of immediate disclosure to a screening regime which we call the Right of Refusal. In this extreme form of a secret reserve the seller never announces the reserve price, she simply accepts or rejects the auction price. We find that the Right of Refusal dominates immediate disclosure if the seller's valuation is a sufficient statistic for the private information of interest. Thus a seller with market-relevant private preference information can benefit from not exercising monopoly price setting power. The result also provides conditions under which a competitive screening equilibrium is more efficient than a signalling mechanism. Broadly speaking, screening is better when the common value aspect in the preferences of the informed and uninformed parties are `aligned', and potential gains from trade to the uninformed party are significant. We believe this conclusion to be of particular interest to the design of privatization schemes.

M.Ing. (Mechanical Engineering) / In this study a manipulator was developed by using system design methods. The main aim was to develop a robot which could be used as a packer in an industrial environment. Pneumatic silinders have been used with great success but up to now seldom in areas where positional accuracy is required. To design a manipulator which is able to operate in a varying work space it must be supplied with specialized sensors. For this reason research was done into image processing. Most of the relevant techniques of image processing arc discussed shortly with particular emphasis on the method of moment invariants for object recognition. Another important aspect covered in the project is the way by which articulated mechanical systems can be modelled kinematically and dynamically. The kinematical description was especially studied and in particular the method introduced by Denavit and Hartenberg and the method of general matrix transformations.