Toru Ishida

JAAMAS Journal 2026 Journal Article

Real-Time Search for Autonomous Agents and Multiagent Systems

• Toru Ishida

Abstract Since real-time search provides an attractive framework for resource-bounded problem solving, this paper extends the framework for autonomous agents and for a multiagent world. To adaptively control search processes, we propose ε-search which allows suboptimal solutions with ε error, and δ-search which balances the tradeoff between exploration and exploitation. We then consider search in uncertain situations, where the goal may change during the course of the search, and propose a moving target search (MTS) algorithm. We also investigate real-time bidirectional search (RTBS) algorithms, where two problem solvers cooperatively achieve a shared goal. Finally, we introduce a new problem solving paradigm, called organizational problem solving, for multiagent systems.

IJCAI Conference 2011 Conference Paper

Constraint Optimization Approach to Context Based Word Selection

• Jun Matsuno
• Toru Ishida

Consistent word selection in machine translation is currently realized by resolving word sense ambiguity through the context of a single sentence or neighboring sentences. However, consistent word selection over the whole article has yet to be achieved. Consistency over the whole article is extremely important when applying machine translation to collectively developed documents like Wikipedia. In this paper, we propose to consider constraints between words in the whole article based on their semantic relatedness and contextual distance. The proposed method is successfully implemented in both statistical and rule-based translators. We evaluate those systems by translating 100 articles in the English Wikipedia into Japanese. The results show that the ratio of appropriate word selection for common nouns increased to around 75% with our method, while it was around 55% without our method.

IJCAI Conference 2009 Conference Paper

• Rie Tanaka
• Yohei Murakami
• Toru Ishida

Machine translation services available on the Web are becoming increasingly popular. However, a pivot translation service is required to realize translations between non-English languages by cascading different translation services via English. As a result, the meaning of words often drifts due to the inconsistency, asymmetry and intransitivity of word selections among translation services. In this paper, we propose context-based coordination to maintain the consistency of word meanings during pivot translation services. First, we propose a method to automatically generate multilingual equivalent terms based on bilingual dictionaries and use generated terms to propagate context among combined translation services. Second, we show a multiagent architecture as one way of implementation, wherein a coordinator agent gathers and propagates context from/to a translation agent. We generated trilingual equivalent noun terms and implemented a Japanese-to-German-and-back translation, cascading into four translation services. The evaluation results showed that the generated terms can cover over 58% of all nouns. The translation quality was improved by 40% for all sentences, and the quality rating for all sentences increased by an average of 0. 47 points on a five-point scale. These results indicate that we can realize consistent pivot translation services through context-based coordination based on existing services.

AAMAS Conference 2008 Conference Paper

Evacuation Guide System based on Massively Multiagent System

• Yuu Nakajima
• Shohei Yamane
• Hiromitsu HATTORI
• Toru Ishida

Using ubiquitous devices such as multifunctional cellular phones and PDAs, we can build a large-scale navigation system for evacuation in the metropolis. Although current navigation systems simply broadcast the same instructions over a large area, the required function is to provide individualized instructions to each user. Our approach is to build a navigation system based on multiagent system which assigns one guide agent to each user. In the system, a guide agent can provide personalized navigation instructions depending on its owner’s surrounding circumstances. We implemented Mega-Navigation system using massively multiagent platform, called Caribbean/Q. Our Mega-Navigation system, which is currently able to work for evacuation in the city of Kyoto, can catch evacuees’ position and offer instructions through GPS-capable cellular phone.

IJCAI Conference 2007 Conference Paper

• Toru Ishida
• Yuu Nakajima
• Yohei Murakami
• Hideyuki Nakanishi

To test large scale socially embedded systems, this paper proposes a multiagent-based participatory design that consists of two steps; 1) participatory simulation, where scenario-guided agents and human-controlled avatars coexist in a shared virtual space and jointly perform simulations, and the extension of the participatory simulation into the 2) augmented experiment, where an experiment is performed in real space by human subjects enhanced by a large scale multiagent simulation. The augmented experiment, proposed in this paper, consist of 1) various sensors to collect the real world activities of human subjects and project them into the virtual space, 2) multiagent simulations to simulate human activities in the virtual space, and 3) communication channels to inform simulation status to human subjects in the real space. To create agent and interaction models incrementally from the participatory design process, we propose the participatory design loop that uses deductive machine learning technologies. Indoor and outdoor augmented experiments have been actually conducted in the city of Kyoto. Both experiments were intended to test new disaster evacuation systems based on mobile phones.

IJCAI Conference 2007 Conference Paper

• Yichuan Jiang
• Toru Ishida

Social law is perceived as evolving through the competition of individual social strategies held by the agents. A strategy with strong authority, accepted by many agents, will tend to diffuse to the remaining agents. The authority of a social strategy is determined by not only the number of but also the collective social positions of its overlaid agents. This paper presents a novel collective strategy diffusion model in agent social law evolution. In the model, social strategies that have strong authority are impressed on the other agents. The agents will accept (partially or in full) or reject them based on their own social strategies and social positions. The diffusion of social strategies proceeds in a series of steps and the final result depends on the interplay between the forces driving diffusion and the counteracting forces.

AAMAS Conference 2007 Conference Paper

Agent Coordination by Trade-off between Locally Diffusion Effects and Socially Structural Influences

• Yichuan Jiang
• Jiuchuan Jiang
• Toru Ishida

There were always two separated methods to make agent coordination: individual-local balance perspective and individualsociety balance perspective. The first method only considered the balance between individual agents and their local neighbors; the second method only considered the balance between individual agents and the whole multi-agent society. However, in reality, the agents will be diffused by their local neighbors as well as influenced by their social contexts simultaneously; therefore, it is necessary to deal with the social performance as well as local performance. To address such problem this paper presents an agent coordination method in an integrative model where we combine the two perspectives together and make trade-off between them. With our presented model, the individual, local and social concerns can be balanced well in a unified and flexible manner. Moreover, the experimental results show that there are often situations in which the two coordination perspectives aren't conflictive but often bring out the better in each other.

AAMAS Conference 2006 Conference Paper

Modeling Agents and Interactions in Agricultural Economics

• Daisuke Torii
• Toru Ishida
• Francois Bousquet

AAMAS Conference 2006 Conference Paper

Multiagent-Based Participatory Design for Ubiquitous Computing

• Toru Ishida

AAMAS Conference 2003 Conference Paper

Can Software Agents Influence Human Relations? - Balance Theory in Agent-mediated Communities -

• Hideyuki Nakanishi
• Satoshi Nakazawa
• Toru Ishida
• Katsuya Takanashi
• Katherine Isbister

AIJ Journal 2003 Journal Article

Controlling the learning process of real-time heuristic search

• Masashi Shimbo
• Toru Ishida

Real-time search provides an attractive framework for intelligent autonomous agents, as it allows us to model an agent's ability to improve its performance through experience. However, the behavior of real-time search agents is far from rational during the learning (convergence) process, in that they fail to balance the efforts to achieve a short-term goal (i. e. , to safely arrive at a goal state in the present problem solving trial) and a long-term goal (to find better solutions through repeated trials). As a remedy, we introduce two techniques for controlling the amount of exploration, both overall and per trial. The weighted real-time search reduces the overall amount of exploration and accelerates convergence. It sacrifices admissibility but provides a nontrivial bound on the converged solution cost. The real-time search with upper bounds insures solution quality in each trial when the state space is undirected. These techniques result in a convergence process more stable compared with that of the Learning Real-Time A ∗ algorithm.

AAMAS Conference 2003 Conference Paper

Scenario Description for Multi-Agent Simulation

• Yohei Murakami
• Tomoyuki Kawasoe
• Reiko Hishiyama
• Toru Ishida

AAAI Conference 1996 Conference Paper

Improving the Learning Efficiencies of Realtime Search

• Toru Ishida

The capability of learning is one of the salient features of realtime search algorithms such as LRTA*. The major impediment is, however, the instability of the solution quality during convergence: (1) they try to find all optimal solutions even after obtaining fairly good solutions, and (2) they tend to move towards unexplored areas thus failing to balance exploration and exploitation. We propose and analyze two new realtime search algorithms to stabilize the convergence process. E-search (weighted realtime search) allows suboptimal solutions with E error to reduce the total amount of learning performed. b-search (reultime search with upper bounds) utilizes the upper bounds of estimated costs, which become available after the problem is solved once. Guided by the upper bounds, S-search can better control the tradeoff between exploration and exploitation.

AAAI Conference 1992 Conference Paper

Moving Target Search with Intelligence

• Toru Ishida

We previously proposed the moving target search (MTS) algorithm, where the location of the goal may change during the course of the search. MTS is the first search algorithm concerned with problem solving in a dynamically changing environment. However, since we constructed the algorithm with the minimum operations necessary for guaranteeing its completeness, the algorithm as proposed is neither efficient nor intelligent. In this paper, we introduce innovative notions created in the area of resource- bounded planning into the formal search algorithm, MTS. Our goal is to improve the efficiency of MTS, while retaining its completeness. Notions that are introduced are (1) commitment to goals, and (2) deliberation for selecting plans. Evaluation results demonstrate that the intelligent MTS is 10 to 20 times more efficient than the original MTS in uncertain situations.

AAAI Conference 1990 Conference Paper

An Organizational Approach to Adaptive Production Systems

• Toru Ishida

Recently-developed techniques have improved the performance of production systems several times over. However, these techniques are not yet adequate for continuous problem solving in a dynamically changing environment. To achieve adaptive real-time performance in such environments, we use an organization of distributed production system agents, rather than a single monolithic production system, to solve problems. Organization seZf-design is performed to satisfy real-time constraints and to adapt to changing resource requirements. When overloaded, individual agents decompose themselves to increase parallelism, and when the load lightens the agents compose with each other to free hardware resources. In addition to increased performance, generalizations of our composition/decomposition approach provide several new directions for organization self-design, a pressing concern in Distributed AI.

AAAI Conference 1988 Conference Paper

Optimizing Rules in Production System Programs

• Toru Ishida

Recently developed production systems enable users to specify an appropriate ordering or a clustering of join operations. Various efficiency heuristics have been used to optimize production rules manually. The problem addressed in this paper is how to automatically determine the best join structure for production system programs. Our algorithm is not to directly apply the efficiency heuristics to programs, but rather to enumerate possible join structures under various constraints. Evaluation results demonstrate this algorithm generates a more efficient program than the one obtained by manual optimization.