Setsuya Kurahashi

Why not multiple solutions: agent-based social interaction analysis via inverse simulation

This paper proposes a new method: inverse simulation for analyzing emergent behaviors of agents in artificial societies, which aims at modeling social interactions in electronic mediated communication. Unlike conventional computational society models, inverse simulation executes simulation steps in the reverse order: set a macro-level objective function, evolve the worlds to fit to the objectives, then observe the micro-level agent characteristics. Genetic algorithms with tabu search attain this. The proposed method is able to optimize multi-modal functions. This means that, from the same initial conditions and the same objective function, we can evolve different results, which we often observe in real world phenomena.

We need multiple solutions for electric equipments configuration in a power plant - applying Bayesian optimization algorithm with Tabu search

This paper addresses electric equipments configuration problems in a power plant. The problems are large-scale nonlinear combinatorial optimization problems with various constraints. In this paper, we apply Bayesian optimization algorithm with Tabu search (Tabu-BOA) to the problems. From intensive experiments, the proposed method has shown the effectiveness to practical problems.

How TRURL Evolves Multiagent Worlds for Social Interaction Analysis

TRURL is an agent-based simulation environment, which is designed to analyze social interactions among agents including software and people in community computing. The unique characteristics of TRURL are summarized as follows: (1) Unlike conventional simulation systems, TRURL has so many predetermined and acquired parameters with which TRURL is able to simulate very complex conditions of the societies. The former parameters have constant values during one simulation cycle, however, the latter parameters change during the interactions. (2) TRURL utilizes Genetic Algorithms to evolve the societies by changing the predetermined parameters to optimize macro-level socio-metric measures. This means TRURL solves large-scale inverse problems. This paper first describes basic principles, architecture, and mechanisms of TRURL. Then, it discusses how TRURL evolves the artificial societies by automated parameters tuning on both micro- and macro-level phenomena grounded in the activities of real worlds.

PATTERN-ORIENTED INVERSE SIMULATION FOR ANALYZING SOCIAL PROBLEMS: FAMILY STRATEGIES IN CIVIL SERVICE EXAMINATION IN IMPERIAL CHINA

This paper proposes an agent-based model to investigate the role of parental relationships and intergenerational reproduction of cultural capital to understand the long-term professional success of an elite family line during the Ming and Qing dynasties in imperial China. We implemented the model by a new method: The pattern-oriented inverse simulation (POIS) method, where multiple patterns observed in an elite family line are employed to guide the model design and test alternative assumptions as family strategies. A genetic algorithm (GA) based inverse technique is applied to fit the simulation outputs with actual data grouped in time intervals as patterns. The simulation results discovered family strategies sustained by the examination systems in imperial China, which relates to important sociological theories on the impact of the intergenerational reproduction of cultural capital within family circles on social inequality in the individual entry in labor and professional markets. The case study also proved that the new methodology of POIS can improve our current practices for systematically exploring simulation parameter space and fit model output with actual data.

The Passenger Decision Making Mechanism of Self-service Kiosk at the Airport

We examine how individuals decide to use self-service technology. The decisions made by individuals between options of service are to be located in various contexts, including that of their traits. We focus on the check-in process for air travelers at the airport and map the actual existing world onto the experimental space to represent the decision making process in an agent-based model (ABM). Real-world data, taken from an airline’s system, is used to verify and validate the model. A cognitive model is implemented in ABM, which utilizes a fuzzy inference system to model each agent’s choice. Passenger behavior is carefully designed based on the knowledge of experienced front-line airport customer-service experts and is also reviewed and clarified by on-site observations. We also discuss how to validate the effectiveness of ABM in the end.

Effects of Collaborative Learning on a Complex Doubly Structured Network

Abstract The purpose of this research is to clarify the actual conditions of understanding of teaching done in a classroom. As a means to do so, we propose a simulation for in-class learning processes with consideration given to academic capability, learning material structure, and collaborative relationships. We build an internal network by estimating the understanding probability network by the use of Item Response Theory (IRT) and estimating the learning material structure model with the use of the Bayesian network. The influence of teaching strategies on learning effects is analysed in the model. Moreover, the influence of the seating arrangement of learners on collaborative learning effects and ability groups are discussed. As a result of the simulation, the following points were found: (1) the learning effects depend on the difference in teaching strategies; (2) a teaching strategy where learning skills, material structure, and collaborative learning are integrated is the most effective; and (3) seating arrangements affects collaborative learning.

Optimising of support plans for new graduate employment market using reinforcement learning

We focused on the problems of the new graduate market in Japan, where the recruitment period starts simultaneously. Therefore the competition among students become fierce, and many students spend a lot of time and efforts for their recruitment activity, but their behaviors are not effective. In order to clarify these problems, we conducted the multi-agented simulation with reinforcement learning. After dividing students into six groups by their ability and aggressiveness, we executed two types of support plans by Actor Critic which is the one of reinforcement learning. Then it was found that the support plans which encourage the students, whose abilities are middle-level and aggressiveness are low-level, are effective to increase final finding employment rate in the recruitment market.

Technology Extraction of Expert Operator Skills from Process Time Series Data

Continuation processes in chemical and/or biotechnical plants always generate a large amount of time series data. However, since conventional process models are described as a set of control models, it is difficult to explain complicated and active plant behaviors. To uncover complex plant behaviors, this paper proposes a new method of developing a process response model from continuous time-series data. The method consists of the following phases: (1) Reciprocal correlation analysis; (2) Process response model; (3) Extraction of control rules; (4) Extraction of a workflow; and (5) Detection of outliers. The main contribution of the research is to establish a method to mine a set of meaningful control rules from a Learning Classifier System using the Minimum Description Length criteria and Tabu search method. The proposed method has been applied to an actual process of a biochemical plant and has shown its validity and effectiveness.

Agent-Based Self-Service Technology Adoption Model for Air-Travelers: Exploring Best Operational Practices

The continuous development of the service economy and an aging society with fewer children is expected to lead to a shortage of workers in the near future. In addition, the growth of the service economy would require service providers to meet various service requirements. In this regard, self-service technology (SST) is a promising alternative to securing labor in both developed and emerging countries. SST is expected to coordinate the controllable productive properties in order to optimize resources and minimize consumer stress. As services are characterized by simultaneity and inseparability, a smoother operation in cooperation with the consumer is required to provide a certain level of service. This study focuses on passenger handling in an airport departure lobby with the objective of optimizing multiple service resources comprising interpersonal service staff and self-service kiosks. Our aim is to elucidate the passenger decision- making mechanism of choosing either interpersonal service or self-service as the check-in option, and to apply it to analyze several scenarios to determine the best practice. The experimental space is studied and an agent-based model is proposed to analyze the operational efficiency via a simulation. We expand on a previous SST adoption model, which is enhanced by introducing the concept of individual traits. We focus on the decision-making of individuals who are neutral toward the service option, by tracking the actual activity of passengers and mapping their behavior into the model. A new method of validation that follows a different approach is proposed to ensure that this model approximates real-world situations. A scenario analysis is then carried out with the aim of exploring the best operational practice to minimize the stress experienced by the air travelers and to meet the business needs of the airline managers at the airport. We collected actual data from the Departure Control System of an airline to map the real-world data to the proposed model. Passenger behavior was extracted by front-line service experts and clarified through consecutive on-site observations.

A Health Policy Simulation Model of Ebola Haemorrhagic Fever and Zika Fever

This study proposes a simulation model of a new type of infectious disease based on Ebola haemorrhagic fever and Zika fever. SIR (Susceptible, Infected, Recovered) model has been widely used to analyse infectious diseases such as influenza, smallpox, bioterrorism, to name a few. On the other hand, Agent-based model begins to spread in recent years. The model enables to represent behaviour of each person in the computer. It also reveals the spread of an infection by simulation of the contact process among people in the model. The study designs a model based on Epstein’s model in which several health policies are decided such as vaccine stocks, antiviral medicine stocks, the number of medical staff to infection control measures and so on. Furthermore, infectious simulation of Ebola haemorrhagic fever and Zika fever, which have not yet any effective vaccine, is also implemented in the model. As results of experiments using the model, it has been found that preventive vaccine, antiviral medicine stocks and the number of medical staff are crucial factors to prevent the spread. In addition, a modern city is vulnerable to Zika fever due to commuting by train.