Save 37% off PRO during our Black Friday Sale! »

First talk in ISAGA summer school 2019

1852ac80648a76e2a64589c7d6ee75c3?s=47 hajimizu
August 19, 2019

First talk in ISAGA summer school 2019

Research on work systems using games and simulations

1852ac80648a76e2a64589c7d6ee75c3?s=128

hajimizu

August 19, 2019
Tweet

Transcript

  1. ISAGA Summer School 2019 Research on work systems using games

    and simulations Aug. 19th, 2019 Hajime Mizuyama mizuyama@ise.aoyama.ac.jp, hajimizu@kth.se 1
  2. ISAGA Summer School 2019 A brief resume skip 2

  3. ISAGA Summer School 2019 Academic background and interests Typical topics

    addressed in MSE Production scheduling and simulation, Production system design and improvement, Process modelling and optimization, etc. Mathematical bases Applied operations research, Applied statistics and machine learning, Discrete event systems, Engineering economics, etc. Recent interests Serious gaming, Participatory simulation, Human computation, Game theory and mechanism design, Multi-agent systems, etc. 3 → shifted/expanded to human aspects
  4. ISAGA Summer School 2019 Agenda • Brief self-introduction • Characterization

    of human-involved work systems • Typical research directions/questions • Combination approach of behavioral and mathematical • Some ongoing cases • Summary 4
  5. ISAGA Summer School 2019 Conventional mechanical view A manufacturing system

    (or a factory) – is composed of machines and human workers, called manufacturing resources, – which carry out manufacturing operations/tasks according to a given plan or schedule. Planning and scheduling – is a function of assigning necessary operations/tasks to manufacturing resources – by an outside decision maker. 5
  6. ISAGA Summer School 2019 Difference b/w machines and humans •

    Humans are inherently autonomous, and it is difficult to completely take planning/scheduling function away from them. • This characteristic becomes relevant especially in an uncertain and changing environment, and the reality is essentially so. • It may work as a source of flexibility, robustness, resilience, etc. of the system, but not always, and how to make it work is not clear. • Further, even machines are made more and more intelligent and autonomous, that is, similar to humans. • Accordingly, the autonomousness may no longer be ignored, and manufacturing systems should be addressed as being composed of autonomous agents, rather than mechanical resources. 6
  7. ISAGA Summer School 2019 Incentivized strategic agent with skills Situational

    and procedural knowledge • Focus and granularity of situational awareness • Forecasting models and decision heuristics What actually are the perceptual and action skills? How to support learning and performing the skills? Preference of one’s own and others’ • Objective function, value function, goals, targets, etc. • Strategic (game-theoretical) interaction with others How to align Incentives towards an overall objective? How to incentivize cooperation/coordination? 7
  8. ISAGA Summer School 2019 Scope of research Systems of interest

    • My focus is expanded from manufacturing systems to a wider range of work systems including service providing systems, supply chains, etc. • They are addressed as a system where humans and machines (or computers) are collaboratively deliver values to customers. Goals and approaches • To deepen the understanding on, and to support enhancing the performance of, such human-involved collaborative work systems. • By a combination approach of behavioral and mathematical. 8
  9. ISAGA Summer School 2019 Agenda • Brief self-introduction • Characterization

    of human-involved work systems • Typical research directions/questions • Combination approach of behavioral and mathematical • Some ongoing cases • Summary 9
  10. ISAGA Summer School 2019 Research directions/questions 10 Skills support/development Cooperation/coordination

    Skills analysis Preference analysis Collaborative work systems
  11. ISAGA Summer School 2019 Example behavioral/mathematical tools 11 Skills support/development

    Pedagogical games Inverse RL Cooperation/coordination Participatory simulation Mechanism design Simulation games Reinforcement learning Skills analysis Experimental economics Game theory Preference analysis Collaborative work systems
  12. ISAGA Summer School 2019 Agenda • Brief self-introduction • Characterization

    of human-involved work systems • Typical research directions/questions • Combination approach of behavioral and mathematical • Some ongoing cases • Summary 12
  13. ISAGA Summer School 2019 Our approach: Still immature and evolving

    What model? • Agent-based model, which first models each of the humans in the system as an incentivized agent with skills, and study the system- level behavior considering the interactions among the agents. Why? • The system’s overall behavior can be related to the actions and their background rationale of each participant. • It will become straightforward to relate serious game and computer simulation. Each player simply corresponds to a computer agent. 13 (a) Construct a mathematical model of the collaboration system
  14. ISAGA Summer School 2019 Model outline: How each agent works

    14 Heuristic/strategy in a dynamic task environment Decision, action Input variables on how current situation is captured The role can be captured as a transformation from input to output. How to define input/output spaces, timing, etc. might also be a key.
  15. ISAGA Summer School 2019 Model outline: How the system rolls

    Humans can observe (some portion of) the system status, and will act on condition of the observation (to achieve their own goals). 15 System status: s0 Observe System status: s1 Intervene System status: s2 System status: s3 System status: s5 Communication among the agents may also be incorporated Each individual is modelled as an incentivized agent with skills
  16. ISAGA Summer School 2019 Our approach: Still immature and evolving

    16 (a) Construct a mathematical model of the collaboration system (d) Collect and analyze the play logs of game (b) (e) Collect and analyze the results of simulation (c) (f) Relate and compare (d) and (e), and draw some findings (b) Develop a serious game based on (a) (c) Develop a computer simulation model based on (a) Refine (a) according to (f) and repeat the analysis Utilize the findings in (f) (f) Relate and compare (d) and (e), and draw some findings
  17. ISAGA Summer School 2019 How to combine two approaches #1

    Behavioral approach • Observe humans’ behavior mainly in the game world Mathematical approach • Need to rely on computer simulation in many cases 17 Hypotheses on the skills Behavioral experiments As-expected and surprising results Computer experiments
  18. ISAGA Summer School 2019 How to combine two approaches #2

    Behavioral approach • Observe humans’ behavior mainly in the game world Mathematical approach • Need to rely on computer simulation in many cases 18 Behavioral experiments Formalize/general ize/quantify Patterns/rules Computer experiments
  19. ISAGA Summer School 2019 Task environment classification: Tentative Single-agent setting

    Multi-agent setting 19 Simple Complex
  20. ISAGA Summer School 2019 Some mathematical bases of modelling Single-agent

    setting Multi-agent setting 20 Simple Complex Advanced RL models, such as DQN Simple MDP and reinforcement learning Periodical optimization Multi-agent reinforcement learning Distributed periodical optimization Game theory models and mechanism design Evolutionary game models
  21. ISAGA Summer School 2019 Agenda • Brief self-introduction • Characterization

    of human-involved work systems • Typical research directions/questions • Combination approach of behavioral and mathematical • Some ongoing cases • Summary 21
  22. ISAGA Summer School 2019 Recent cases Single-agent setting Multi-agent setting

    22 Simple Complex Market research, and Auction-based scheduling In-house supply chain of a large steel manufacturer (ColPMan) Food supply chain (How to reduce food loss/waste?) Operations in a restaurant floor Crane operation task in a factory shop floor Knowledge sharing in a sales department
  23. ISAGA Summer School 2019 In-house supply chain of a manufacturer

    23 HQ Customers Downstream factories Upstream factory Order materials Assign orders Order Delivery Various decisions concerning production management are distributed to the sites. How to collaboratively address stationary and non-stationary variations?
  24. ISAGA Summer School 2019 ColPMan game 24

  25. ISAGA Summer School 2019 Game flow #1 Time Term 1

    Term 2 Term 3 ... Period 1-5 Period 1-5 Period 1-5 ... A team of players Simulation Simulation Simulation Simulation Planning information Progress information
  26. ISAGA Summer School 2019 Discrete event simulation representing SC operations

    according to given plans under various disturbances Game flow #2 Distributed planning with controlled communication DSF1 player DSF2 player DSF3 player USF player HQ player USF DSF3 DSF1 DSF2 HQ Planning information Progress information
  27. ISAGA Summer School 2019 How this case is (still being)

    approached 1st Round • ConPMan game (ver. 1.0), which uses table discussion, is created. • Discussion data are analyzed exploratorily by protocol analysis. 2nd Round • ConPMan game (ver. 2.0), which uses online chat, is created. • What information is shared, when, with whom is classified. 3rd Round • Planning decisions in each classified information sharing types is mathematically formulated as distributed optimization. • Which type leads to high performance depends on the environment. 27
  28. ISAGA Summer School 2019 How this case is (still being)

    approached 1st Round 2nd Round 3rd Round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
  29. ISAGA Summer School 2019 How combined approach has been used

    Behavioral approach • Multi-player computer game with a communication channel. Mathematical approach • Mathematical optimization divided into sub-problems. 29 Behavioral experiments Formalize/general ize/quantify Patterns/rules Computer experiments
  30. ISAGA Summer School 2019 Crane operation skills example In a

    factory, different kinds of jobs arrive randomly, and they are expected to be finished by their assigned due dates. An operator needs to handle the jobs with a crane between multiple LIFO buffers and a machine. High production rate and low due date lateness penalty should be accomplished. The performance depends heavily on the skills of the operator. 30
  31. ISAGA Summer School 2019 How combined approach has been used

    Behavioral approach • Single-player computer game Mathematical approach • Simple agent-based simulation • Reinforcement learning (under construction) 31 Behavioral experiments Formalize/general ize/quantify Patterns/rules Computer experiments
  32. ISAGA Summer School 2019 Waitperson skills example #1 32 Waiting

    Room Dining Hall
  33. ISAGA Summer School 2019 Waitperson skills example #2 33 •

    Six tables (squares) are handled by a single waitperson (a blue circle). • The color of each table and dish changes from white to yellow, orange, and finally red as waiting time passes.
  34. ISAGA Summer School 2019 How combined approach has been used

    Behavioral approach • Single-player computer game • Multi-player computer game (under construction) Mathematical approach • Simple agent-based simulation • Reinforcement learning (under construction) 34 Behavioral experiments Formalize/general ize/quantify Patterns/rules Computer experiments
  35. ISAGA Summer School 2019 Milk supply chain and one-third rule

    35 One-third rule Production date Expiration date Delivery limit Sales limit Some argue that this rule is a main cause of the huge food loss and others believe it enhances the supply chain efficiency. Field experiments yield contradictory and inconclusive results.
  36. ISAGA Summer School 2019 How combined approach has been used

    Behavioral approach • Multi-player computer game with auction mechanism. Mathematical approach • Evolutionary game simulation with logit choice models (under construction). 36 Behavioral experiments Patterns/rules Computer experiments
  37. ISAGA Summer School 2019 Agenda • Brief self-introduction • Characterization

    of human-involved work systems • Typical research directions/questions • Combination approach of behavioral and mathematical • Some ongoing cases • Summary 37
  38. ISAGA Summer School 2019 Summary • I am interested in

    human involved collaborative work systems, and addressing them through a combination approach/perspective of behavioral and mathematical. • The combination approach is characterized by agent-based modelling framework, and how each agent is captured (i.e. as an incentivized agent with skills). • This approach is expected to ultimately reveal how autonomous human workers in a work system should be supported so as to make the most of their potentials. • My next talk on Wednesday will be on “capturing, analyzing and interpreting data from games.” 38
  39. ISAGA Summer School 2019 Thank you! Questions & Comments are

    welcome