Infinite-Horizon Proactive Dynamic DCOPs Khoi Hoang Ferdinando - PowerPoint PPT Presentation

Infinite-Horizon Proactive Dynamic DCOPs Khoi Hoang Ferdinando Fioretto Ping Hou William Yeoh Roie Zivan Makoto Yokoo New Mexico State University All About Discovery! nmsu.edu New Mexico State University

Outline Ø Distributed Constraint Optimization Problems Ø Dynamic DCOPs Ø Proactive Dynamic DCOPs Ø Infinite-Horizon Proactive Dynamic DCOPs * Ø Overview and Details New Mexico State University All About Discovery! nmsu.edu

Outline Ø Distributed Constraint Optimization Problems Ø Dynamic DCOPs Ø Proactive Dynamic DCOPs Ø Infinite-Horizon Proactive Dynamic DCOPs * Ø Overview and Details New Mexico State University All About Discovery! nmsu.edu

Outline Ø Distributed Constraint Optimization Problems Ø Dynamic DCOPs Ø Proactive Dynamic DCOPs Ø Infinite-Horizon Proactive Dynamic DCOPs * Ø Overview and Details New Mexico State University All About Discovery! nmsu.edu

Outline Ø Distributed Constraint Optimization Problems Ø Dynamic DCOPs Ø Proactive Dynamic DCOPs Ø Infinite-Horizon Proactive Dynamic DCOPs * Ø Overview and Details New Mexico State University All About Discovery! nmsu.edu

Outline Ø Distributed Constraint Optimization Problems Ø Dynamic DCOPs Ø Proactive Dynamic DCOPs Ø Infinite-Horizon Proactive Dynamic DCOPs * Ø Overview and Details New Mexico State University All About Discovery! nmsu.edu

Distributed Constraint Optimization Problems A B C New Mexico State University All About Discovery! [1] Modi et al., ADOPT: Asynchronous Distributed Constraint Optimization with Quality Guarantees, 2005 nmsu.edu

Distributed Constraint Optimization Problems A x A x B f AB (x A ,x B ) f AB 0 0 5 0 1 10 … … … B C New Mexico State University All About Discovery! [1] Modi et al., ADOPT: Asynchronous Distributed Constraint Optimization with Quality Guarantees, 2005 nmsu.edu

Distributed Constraint Optimization Problems A x A x B f AB (x A ,x B ) x C x A f CA (x C ,x A ) f AB f CA 0 0 5 0 0 7 0 1 10 0 1 4 … … … … … … B C f BC x B x C f BC (x B ,x C ) 0 0 3 0 1 12 … … … New Mexico State University All About Discovery! [1] Modi et al., ADOPT: Asynchronous Distributed Constraint Optimization with Quality Guarantees, 2005 nmsu.edu

Distributed Constraint Optimization Problems x A = ? A B C x B = ? x C = ? Maximize f AB + f BC + f CA New Mexico State University All About Discovery! [1] Modi et al., ADOPT: Asynchronous Distributed Constraint Optimization with Quality Guarantees, 2005 nmsu.edu

Distributed Constraint Optimization Problems • Meeting scheduling problems • Smart devices scheduling • Resource allocation • Sensor network New Mexico State University All About Discovery! nmsu.edu

Limitations • DCOPs – Static problem – Not consider possible changes New Mexico State University All About Discovery! nmsu.edu

Limitations • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 React New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 X 0 New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 P 1 X 0 New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 P 1 X 0 React New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 P 1 X 0 X 1 New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 P 1 P 2 X 0 X 1 New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 P 1 P 2 X 0 X 1 React New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Dynamic DCOPs • DCOPs – Static problem – Not consider possible changes • Dynamic DCOPs – Reacting to changes of the problem P 0 P 1 P 2 X 0 X 1 X 2 New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Limitations • Dynamic DCOPs – Not take advantage of possible changes – Good for current, bad for future (myopic solutions) New Mexico State University All About Discovery! [2] R. Lass et al., Dynamic distributed constraint reasoning, 2005 nmsu.edu [3] A. Petcu and B. Faltings, Superstabilizing, fault-containing multiagent combinatorial optimization, 2005

Limitations • Dynamic DCOPs – Not take advantage of possible changes – Good for current, bad for future (myopic solutions) • How about if we know – How often the problems change – Knowledge about possible changes New Mexico State University All About Discovery! nmsu.edu

Proactive Dynamic DCOPs • Knowledge about changes of random events – Initial distribution and transition function • Solve all the problems beforehand up to horizon h • Keep the solution at time step h P 0 P 1 P h New Mexico State University All About Discovery! [4] Hoang et al., Proactive Dynamic Distributed Constraint Optimization, 2016 nmsu.edu

Proactive Dynamic DCOPs • Knowledge about changes of random events – Initial distribution and transition function • Solve all the problems beforehand up to horizon h • Keep the solution at time step h P 0 P 1 P h X 0 X 1 X h Proactive New Mexico State University All About Discovery! [4] Hoang et al., Proactive Dynamic Distributed Constraint Optimization, 2016 nmsu.edu

Limitations Is this solution optimal from h onwards??? P 0 P 1 P h X 0 X 1 X h Proactive New Mexico State University All About Discovery! [4] Hoang et al., Proactive Dynamic Distributed Constraint Optimization, 2016 nmsu.edu

Key contributions • Infinite-Horizon Proactive Dynamic DCOPs – Optimal solution from h onwards – Based on converged distribution at h* – Proactive vs. Reactive dynamic DCOP algorithms (first time!!!) P 0 P 1 P h P h* Infinite-Horizon Proactive New Mexico State University All About Discovery! nmsu.edu

Key contributions • Infinite-Horizon Proactive Dynamic DCOPs – Optimal solution from h onwards – Based on converged distribution at h* – Proactive vs. Reactive dynamic DCOP algorithms (first time!!!) P 0 P 1 P h P h* X h* Infinite-Horizon Proactive New Mexico State University All About Discovery! nmsu.edu

Key contributions • Infinite-Horizon Proactive Dynamic DCOPs – Optimal solution from h onwards – Based on converged distribution at h* – Proactive vs. Reactive dynamic DCOP algorithms (first time!!!) P 0 P 1 P h P h* X 0 X 1 X h* X h* Infinite-Horizon Proactive New Mexico State University All About Discovery! nmsu.edu

Content Ø Distributed Constraint Optimization Problem • Proactive Dynamic DCOPs • Infinite-Horizon Proactive Dynamic DCOPs • Algorithms • Experiments • Conclusions New Mexico State University All About Discovery! nmsu.edu