| 研究課題/領域番号 |
15J07801
|
|---|
| 研究機関 | 東京大学 |
|---|
研究代表者 |
顧 ジェ 東京大学, 新領域創成科学研究科, 特別研究員(DC2)
|
|---|
| 研究期間 (年度) |
2015-04-24 – 2017-03-31
|
| キーワード | knowledge management / agent-based modeling / behavioral experiment / organizational theory / structure and network / decision making / incentive policy / bounded rationality |
|---|
| 研究実績の概要 |
Work done: A knowledge management (KM) has been designed; a conceptual framework of the game has been established with a detailed specification of agent, organization, and environment; the game has been implemented in both agent-based simulation and behavioral experiment; the agent-based model has been refined and verified; a long-term macroscopic non-monotonicity in organizational performance alongside the improvement of network accessibility has been identified; the cause has been elucidated which is related to the turbulence of the environment; a gaming software has been developed and the behavioral experiments have been conducted; empirical data has been collected; results comparison between simulation and experiment show a strong violation of perfect rationality and reveal a scarcity heuristic in human agents repeated decision making; based on the empirical evidence, the strategy attraction rule of agent-based model has been modified and enhanced; new results demonstrate a better fit than before. Goal achieved: The development of new integrated methodology with agent-based simulation and behavioral experiments has been achieved. The impact of microscopic agent’s KM effort on macroscopic organizational outcomes and various causal relationships have been elucidated. Organizational policies and interventions have been evaluated. Recognition and Publications: a book chapter, two journal papers, and six full conference papers have been published. A graduate scholar award and a nomination of the best student presentation have been given as excellent performance recognition.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
The work planned has been exceedingly difficult since little literature has paved the way ahead and growing related disciplines/research fields need to be explored and investigated in details. Conducting controlled human behavioral experiments, especially on incentive effectiveness, is extremely challenging and cost risky. With passion and hard work, the progress and timeline can be under control. However, with lessons learnt and experience gained, more efficient time management and effective outcome production are expected in the future work.
|
| 今後の研究の推進方策 |
The further work will mainly focus on introducing incentive policy as organizational control and co-evolution adaptation features on top of the previous framework and model. To be specific, the following work will be carried out. To establish a conceptual framework for agent’s behavioral decision making under social and economic dilemma (2016.4); To develop a gaming software for conducting the behavioral experiments with human participants (2016.5); To conduct a series of behavioral experiment to validate and refine the conceptual framework and collect empirical data for knowledge management policies design and parameters setting (2016.6); To incorporate a social and economic dilemma to the agent-based model design and conduct simulation experimentation (2016.7); To compare both simulation and experiments results and fact-find the causes of accordance or discordance, propose further improvement and testing (2016.8); To submit doctoral course thesis for graduation and prepare for attending international conference (2016.8-9); To consolidate previous work into a journal paper for publication (2016.9-10); To build additional features and extend the knowledge management framework with multi-faceted indicators for evaluation (2016.10-11); To introduce an administrative role to the framework and incorporate co-evolution mechanism with member agents (2016.12); To implement the organizational knowledge management into both agent-based simulation and behavioral experiments (2017.1-2); Prepare a journal paper for publication (2017.3)
|
