This volume contains 13 thoroughly refereed and revised papers detailing recent advances in research on trading agents, negotiating agents, dynamic pricing, and auctions. They were originally presented at the 10th International Workshop on Agent-Mediated Electronic Commerce (AMEC 2008) collocated with AAMAS 2008 in Estoril, Portugal, or the 6th Workshop on Trading Agent Design and Analysis (TADA 2008) collocated with AAAI 2008 in Chicago, IL, USA. The papers originating from AMEC 2008 address agent modeling and multi-agent problems in the context of e-negotiations and e-commerce. The TADA papers stem from the effort to design scenarios where trading agents and market designers can be pitched against each other in applications from supply chain management and procurement. They are all characterized by interdisciplinary research combining fields such as artificial intelligence, distributed systems, game theory, and economics.

The design and analysis of electronic commerce systems in which agents are - ployed involves finding solutions to a large and diverse array of problems, concerning individual agent behaviors, interact

The use of simple auction mechanisms like the GSP in online advertising can lead to significant loss of efficiency and revenue when advertisers have rich preferences — even simple forms of expressiveness like budget constraints can lead to suboptimal outcomes. While the optimal allocation of inventory can provide greater efficiency and revenue, natural formulations of the underlying optimization problems grow exponentially in the number of features of interest, presenting a key practical challenge. To address this problem, we propose a means for automatically partitioning inventory into abstract channels so that the least relevant features are ignored. Our approach, based on LP/MIP column and constraint generation, dramatically reduces the size of the problem, thus rendering optimization computationally feasible at practical scales. Our algorithms allow for principled tradeoffs between tractability and solution quality. Numerical experiments demonstrate the computational practicality of our approach as well as the quality of the resulting abstractions.

AAAI was pleased to present the AAAI‐08 Workshop Program, held Sunday and Monday, July 13–14, in Chicago, Illinois, USA. The program included the following 15 workshops: Advancements in POMDP Solvers; AI Education Workshop Colloquium; Coordination, Organizations, Institutions, and Norms in Agent Systems, Enhanced Messaging; Human Implications of Human‐Robot Interaction; Intelligent Techniques for Web Personalization and Recommender Systems; Metareasoning: Thinking about Thinking; Multidisciplinary Workshop on Advances in Preference Handling; Search in Artificial Intelligence and Robotics; Spatial and Temporal Reasoning; Trading Agent Design and Analysis; Transfer Learning for Complex Tasks; What Went Wrong and Why: Lessons from AI Research and Applications; and Wikipedia and Artificial Intelligence: An Evolving Synergy.

Objectives/Hypothesis: The perichondrial cutaneous graft (PCCG) possesses unique characteristics that make them a propitious reconstructive option in facial plastic surgery. The PCCG is easily harvested from the conchal bowl. Notable characteristics are that it does not contract, unexcelled color match for resurfacing facial skin, and minimal donor site morbidity. This free graft frequently offers an expeditious solution to an otherwise more complicated reconstructive effort.Study Design: Retrospective review of an academic otolaryngology and facial plastic surgery practice.Methods: Patients requiring head and neck reconstruction for cutaneous deficiencies were studied. The PCCG is harvested from the anterior conchal bowl. This is technically easy, and the perichondrium. is tightly adherent to the dermis in this area. The donor site is closed by resecting a fenestra of conchal cartilage and rotating a posterior auricular interpolated island flap into the defect (the "flip-flop-flap"). The posterior auricular defect is easily closed in a linear fashion. The main outcome measures were cosmetic result, graft survival, donor site morbidity, and complications.Results: There are 406 PCCGs in our series. Patients ranged in age from 7 days to 94 years. There were 170 grafts used for trauma and 236 used for reconstruction after skin cancer resection. Over the past 30 years in observing these grafts, there are no contractions noted in infants and children, growth with maturity is noteworthy. Cosmesis is excellent and in most cases superior to other skin grafting techniques. We have had four total failures and six partial losses of less than 30% for the PCCG. All complete failures were in patients with a smoking history.Conclusions: The PCCG is a very reliable flap for reconstruction of facial defects. It has been used in elderly and heavy smoking patients with minimal complications, attesting to its viability. The graft provides excellent cosmesis and it is an expeditious alternative to commonly used local flaps. It is especially useful in pediatric patients because the graft actually expands with growth. This is in contrast to the disadvantages of split thickness and full thickness skin grafts that predictably contract with maturity. Local flaps often lack adequate laxity for common implementation and make the PCCG a propitious choice in the pediatric patient.

Single-good ascending auctions, including the English Auction and its close variants (e.g. eBay), are the most widely used type of auction. Hence, effective strategies for such auctions can have an enormous economic impact. To maximize profit, a seller should try to set a reserve price high enough to extract the highest bidder’s full value without blocking the bidder out altogether. In isolated auctions where bidders have static, independent, identically distributed, private values from a known distribution, it is wellunderstood how to compute the optimal reserve price. However, these assumptions rarely hold in practice. First, the value distributions are not known ex ante. Second, auctions rarely run in isolation [2, 7]. Third, value distributions are typically non-stationary. Estimating the value distribution from bids is non-obvious and non-trivial because the distribution of bids is not the same as the distribution of values [4, 6]. Jiang and LeytonBrown [6] addressed how “hidden bids” in online auctions can skew the bidding distribution away from the underlying value distribution. They were able to effectively infer the value distribution when given the parametrized form. Haile and Tamer’s [4] method infers bounds on the distribution, making no a priori assumptions about the bidder valuations and very minimal assumptions about bidder behavior. However, their approach was not complete, as it did not provide a way to choose the reserve price within the bounds. Additionally, neither of the aforementioned approaches addressed the issue of non-isolated auctions or non-stationary distributions. Although Juda and Parkes [7] generalize Haile and Tamer’s work somewhat to allow for bidders that participate in multiple auctions, and Gerding et al. [2] analyzed reserve pricing in the presence of multiple competing auctions, those models are simplifications. Indeed, performing a full game-theoretic analysis of setting reserve prices in a dynamic real-world context is prohibitively complex. We present an automated methodology and system for computing reserve prices for real-world ascending auctions. Our initial system is based on the approach of Haile and Tamer [4], but with the addition of our own technique for

We present the design of a banner advertising auction which is considerably more expressive than current designs. We describe a general model of expressive ad contract/bidding and an allocation model that can be executed in real time through the assignment of fractions of relevant ad channels to specific advertiser contracts. The uncertainty in channel supply and demand is addresscd by the formulation of a stochastic combinatorial optimization problem for channel allocation that is rerun periodically. We solve this in two different ways: fast deterministic optimization with respect to expectations; and a novel online sample-based stochastic optimization method-- that can be applied to continuous decision spaces--which exploits the deterministic optimization as a black box. Experiments demonstrate the importance of expressive bidding and the value of stochastic optimization.

In human space exploration missions, there will be a need to provide voice communications services. In this work we focus on the performance of Voice over IP (VoIP) techniques applied to space networks, where long range latencies, simplex links, and significant bit error rates occur. Link layer and network layer overhead issues are examined. We posit that imposing additional speech processing latencies in the form of multiple frames per packet is tolerable in the space regime, and show resulting performance overhead improvements. Furthermore, we find that even with channel bit error rates of 10 -5 and 10 -4 , the frame size does not severely degrade the original speech.

We study autonomic resource allocation among multiple applications based on optimizing the sum of utility for each application. We compare two methodologies for estimating the utility of resources: a queuing-theoretic performance model and model-free reinforcement learning. We evaluate them empirically in a distributed prototype data center and highlight tradeoffs between the two methods

Autonomic (self-managing) computing systems face the critical problem of resource allocation to different computing elements. Adopting a recent model, we view the problem of provisioning resources as involving utility elicitation and optimization to allocate resources given imprecise utility information. In this paper, we propose a new algorithm for regret-based optimization that performs significantly faster than that proposed in earlier work. We also explore new regret-based elicitation heuristics that are able to find near-optimal allocations while requiring a very small amount of utility information from the distributed computing elements. Since regret-computation is intensive, we compare these to the more tractable Nelder-Mead optimization technique w.r.t. amount of utility information required.

A major goal of autonomic computing is to dynamically allocate computational resources so as to continually optimize high-level policy objectives. A key challenge to achieving this goal is to accurately estimate the impact of resource-level changes on application performance with respect to a Service Level Ageement (SLA). We compare two methodologies for accomplishing this: (i) developing a queuing-theoretic performance model for an application, and fitting its parameters online based on current state; (ii) using modelfree reinforcement learning of resource valuation estimates based on trial-and-error learning. We describe these approaches in the context of a distributed architecture in which servers are allocated amongst multiple applications with independent time-varying loads. Each application has a local utility function, based on SLA payments as a function of relevant performance metrics. The overall system goal is to maximize the sum of local utility functions. Individual applications use one of the above methodologies to estimate resource valuations, which are then used by a resource arbiter to compute optimal allocations. We present empirical data illustrating the practicality and effectiveness of both methods in a realistic data center prototype. We highlight important tradeoffs between the methods, and point out potential benefits of a hybrid approach combining both methods.

The characteristics of various IP-based satellite network architectures are explored in order to assess how best to apply IP multicast and reliable multicast building blocks. An approach is formulated that is adapted to the satellite link layers, error and outage patterns.