Naoki Fukuta

BiddingBot: a multiagent support system for cooperative bidding in multiple auctions

Online auctions are becoming an increasingly important channel for electronic commerce. There exist more than 150 online auction sites on the Internet. It is difficult for users to attend, monitor, and bid at multiple auction sites simultaneously. The authors propose BiddingBot which is a multi agent system that supports users in attending, monitoring, and bidding in multiple auctions. BiddingBot monitors prices of goods in several online auction sites to get reasonable market prices of goods, and uses a new cooperative bidding mechanism to effectively bid in auctions.

Towards Better Approximation of Winner Determination for Combinatorial Auctions with Large Number of Bids

We propose new approximate algorithms for combinatorial auctions with massively large number of (more than 100,000) bids. In this paper, we focus on a more practical approximated algorithm in the context of revenue maximization. We propose a hill-climbing greedy algorithm, a SA-like random search algorithm, and their enhancement for searching multiple key parameter values. The experimental results demonstrate that our algorithms perform approximately 0.997 optimality compared with the optimal solutions and better than previously presented approximated algorithms. We also demonstrate that our algorithms are a kind of anytime algorithm that bring better results in shorter computational time that can be applied to large and dynamic electronic markets.

Fine-grained efficient resource allocation using approximated combinatorial auctions: A parallel greedy winner approximation for large-scale problems

Combinatorial auctions, one of the most popular market mechanisms, have a huge effect on electronic markets and political strategies. Combinatorial auctions provide suitable mechanisms for efficient allocation of resources to self-interested attendees. On the other hand, efficient resource allocation is also becoming crucial in many computer systems that should manage resources efficiently. Considering ubiquitous computing scenarios, the ability to complete an auction within a fine-grained time period without loss of allocation efficiency is in strong demand. Furthermore, to achieve such scenarios, it is very important to handle a large number of bids in an auction. In general, the optimal winner determination problem of a combinatorial auction is NP-hard. Thus, much work focuses on tackling the computational costs for winner determination. In this paper, we show that our approximation algorithms provide sufficient quality of winners for auctions that have a large number of bids on hard time constraints. Furthermore, we compare and discuss desirable properties of such approximation algorithms to be embedded in application systems.

Periodical Resource Allocation Using Approximated Combinatorial Auctions

Self-healing in fault tolerant multi-agent systems is the system ability to automatically detect, diagnose, and repair the faults. However, most of the available solutions are fragile in incomplete, uncertain, and dynamic situations. This paper proposes a novel economically inspired self-healing model for fault tolerant Multi-Agent Systems where the agents are self-interested autonomic elements collaborate to achieve fault tolerance as a given high-level objective of the system. It is an effective solution for dynamic situations with a high possibility of uncertainty. The proposed model is in fact toward responding the challenge of negotiation theory for autonomic systems introduced by IBM. In particular, it is an inspiration of general explanation of Communism, Socialism, and Capitalism. Extensive experiments illustrate the effectiveness of the proposed social approach in comparison to the cases of no help situation, using purely redundant components, and helping without using a social value.Combinatorial auction, one of most popular market mechanisms, is well-known mechanism for effective resource allocation to self-interested agents. In real scenarios, since the auction mechanism has to clear the market very frequently, actually there is not enough time to compute exact optimal winners. In this paper, we show that some approximation algorithms provide sufficient quality of winners for auctions that have large number of bids but have hard time constraints. Furthermore, we compare and discuss about desirable properties of such approximation algorithms to be embedded in application systems.

An experimental analysis of biased parallel greedy approximation for combinatorial auctions

Considering emerging demands for auction based efficient resource allocations, the ability to complete an auction within a fine-grained time period without loss of allocation efficiency is in strong demand. Recently, an algorithm has been proposed to obtain near-optimal solutions for winner determination problem on combinatorial auctions within a very short computation time. However, it is demanded to analyse and clarify the main factor of such an excellent performance of the algorithm. Also, for practical use, there are demands to clarify the actual implementation-level performance of the algorithm compared to major commercial-level generic problem solvers. In this paper, we show an analysis about performance issues of biased parallel greedy updating approach in various experimental settings. Furthermore, we show that the algorithm has a certain advantage to solve a kind of large-size high-complexity problems when it is compared to a latest commercial-level implementation of generic LP solver through various experiments.

DODDLE-OWL: Interactive Domain Ontology Development with Open Source Software in Java

In this paper, we propose an interactive domain ontology development environment called DODDLE-OWL. DODDLE-OWL refers to existing ontologies and supports the semi-automatic construction of taxonomic and other relationships in domain ontologies from documents. Integrating several modules, DODDLE-OWL is a practical and interactive domain ontology development environment. In order to evaluate the efficiency of DODDLE-OWL, we compared DODDLE-OWL with popular manual-building method. In order to evaluate the scalability of DODDLE-OWL, we constructed a large sized ontology over 34,000 concepts in the field of rocket operation using DODDLE-OWL. Through the above evaluation, we confirmed the efficiency and the scalability of DODDLE-OWL. Currently, DODDLE-OWL is open source software in Java and has 100 and more users from 20 and more countries.

Short-time approximation on combinatorial auctions: a comparison on approximated winner determination algorithms

Combinatorial auctions, one of the most popular market mechanisms, have a huge effect on electronic markets and political strategies. In real scenarios, since the auction mechanism has to clear the market very frequently, actually there is not enough time to compute exact optimal winners. Some works try to achieve approximate solutions in winner determination. In this paper, we compare four approximation algorithms, and reveal how those algorithms work better on which settings. Especially, we focused on comparing them in the setting of that has many number of bids but only a limited time period can be used for each approximation.

DODDLE-OWL: a domain ontology construction tool with OWL

In this paper, we propose a domain ontology construction tool with OWL The advantage of our tool is focusing the quality refinement phase of ontology construction Through interactive support for refining the initial ontology, OWL-Lite level ontology, which consists of taxonomic relationships (defined as classes) and non-taxonomic relationships (defined as properties), is constructed effectively The tool also provides semi-automatic generation of the initial ontology using domain specific documents and general ontologies.

MiNet: Building Ad-Hoc Peer-to-Peer Networks for Information Sharing Based on Mobile Agents

The Internet is a very popular for information sharing technology since users can share information in organizations and communities. In this paper, we present a flexible peer-to-peer networking technology for information sharing on the Internet called MiNet. In some certain communities, MiNet can construct an ad-hoc network for information sharing. MiNet enables users to share information based on mobile agents, which are implemented in a mobile agent framework MiLog. MiNet can construct ad-hoc peer-to-peer networks by encapsulating information and sending it as mobile agents that can migrate in MiNet beyond firewalls, proxies, and NATs in LANs. Therefore, MiNet can construct VPNs, which consist of several LANs covered by firewalls, etc. MiNet agents can automatically choose a destination platform according to its policies. We show the document sharing system MiDoc as an application based on MiNet. Since MiDoc is implemented using MiNet, MiDoc users can share any document among any LANs.

Towards On-the-Fly Ontology Construction – Focusing on Ontology Quality Improvement

In order to realize the on-the-fly ontology construction for the Semantic Web, this paper proposes DODDLE-R, a support environment for user-centered ontology development. It consists of two main parts: pre-processing part and quality improvement part. Pre-processing part generates a prototype ontology semi-automatically, and quality improvement part supports the refinement of it interactively. As we believe that careful construction of ontologies from preliminary phase is more efficient than attempting generate ontologies full-automatically (it may cause too many modification by hand), quality improvement part plays significant role in DODDLE-R. Through interactive support for improving the quality of prototype ontology, OWL-Lite level ontology, which consists of taxonomic relationships (class – sub class relationship) and non-taxonomic relationships (defined as property), is constructed efficiently.