the workflow framework for metaheuristic search on grid environment

40
The Workflow Framework for Metaheuristic Search on Grid Environment Dang-Khoa Tran Van Hoai Tran Faculty of Computer Science and Engineering HCMC University of Technology

Upload: brygid

Post on 23-Feb-2016

57 views

Category:

Documents


0 download

DESCRIPTION

The Workflow Framework for Metaheuristic Search on Grid Environment. Dang-Khoa Tran Van Hoai Tran Faculty of Computer Science and Engineering HCMC University of Technology. Metaheuristic. Combinatorial optimization problems are often NP-hard, cannot solve in reasonable time. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: The Workflow Framework for Metaheuristic  Search  on Grid Environment

The Workflow Framework for Metaheuristic Search on

Grid EnvironmentDang-Khoa Tran

Van Hoai TranFaculty of Computer Science and Engineering

HCMC University of Technology

Page 2: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 2

Combinatorial optimization problems are often NP-hard, cannot solve in reasonable time.

Meta-heuristic is not like the exact methods, it can find reasonable solutions in acceptable time.

Metaheuristic

Page 3: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 3

Improving the solutions by expanding the exploration space and with various strategies◦ Take more time◦ … or required more computing resources

Our approach: using Grid Computing to solve the metaheuristic problem.

Metaheuristic

Page 4: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 4

Meta-heuristic and Grid Computing Design goals The framework designs Experiments and results

Content

Page 5: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 5

Grid Computing is the parallel and distributed system that enables the sharing and using resources easily.

Grid programming paradigm: using Parallel Object C++ (POP-C++) for application development.◦ Similar to C++ with a few keyword.◦ The objects are automatically created and

executed on Grid.

Grid computing

Page 6: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 6

User is usually not familiar with Grid programming model.

Application testing and debugging on Grid is difficult.

Make use of Grid computing by executing many search algorithm concurrently.

Some challenges

Page 7: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 7

Provide workflow model to solve the problems

Easy to develop applications◦ Using OO programming◦ Hide the Grid programming model from user

Test and debug the applications Deploy and execute on Grid environment

easily Running efficiency

Design Goals for the Framework

Page 8: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 8

Parallel Metaheuristic Library

Library Architecture

Search Algorithms

Workflow Engine

POP-C++ Runtime

Computing Resources(Workstations / Clusters / Grid)

Applications

Page 9: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 9

Using the search algorithms◦ Inherits from predefined interfaces◦ Implements the abstract methods

Using workflow◦ Create dependency graph◦ The workflow engine will handles the execution

Easy to Use+fullEvaluate() : double+Serialize() : void+printOn() : void+operator =()+operator ==() : bool

edaSolution

mySolution

Page 10: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 10

// Define search algorithm myTwoOpt two_opt(graph); myTwoOptNext two_opt_next(graph); edaBestImprSelect moveSelect; edaHC hcSearch( &two_opt, &two_opt_next, &moveSelect); /* Define more search algorithms... */ // Define workflow control: Using Parallel workflow edaParWrapperControl sfControl; // Create dependency graph int hcId = sfControl.insertVertex( &hcSearch ); sfControl.insertEdge( hcId, saId ); // Do search sfControl.search(route)

Easy to Use – Example

Page 11: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 11

Sample workflow

HC

TS

SA

SA TS

SA

HCS

Initial solution

Page 12: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 12

Easy to Deploy

Page 13: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 13

Testing and Debugging◦ Using the sequential workflow◦ Test and debug like normal applications

Easy to Test and Debug

Page 14: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 14

The Traveling Salesman Problem

Page 15: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 15

The Traveling Salesman Problem Workflow design for testing

HC

SA

TS

HC

HC

SA

TS

HC

HC...HC

Input: graph Output: path

Page 16: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 16

The Traveling Salesman Problem

Graph/Branches

1 5 10

PCB 442 55845 54778 54171ALI 535 2136 2140 2250EIL 101 675 679 675PR 1001 293180 288059 288059

Page 17: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 17

Logic Circuit Optimization

• Size optimization• Delay optimization• Energy optimization• …

Page 18: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 18

Workflow design

Logic Circuit Optimization

HC

TS SA TS SA

HCTS SA TS SA

TS SA TS SA

Page 19: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 19

Logic Circuit Optimization Sample circuit: IWLS93 Sequential execution (1 branch of the workflow) Literal optimization

Circuit #lit(in) MIS RAMBO Pert/Sim LogicOptC3540 2778 2145 1976 1876 1548

C7552 6583 3513 3521 2851 2125

C6288 7184 5237 4666 3766 3490

des 7657 6095 6145 5718 3664

apex6 1422 1260 1291 1086 788

Page 20: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 20

Logic Circuit Optimization Execute concurrently on Grid Execute a few times each (average results) Literal optimization

Circuit 1 branch 5 braches 10 branchesC3540 1548 1431 1388

C7552 2125 2149 2074

C6288 3490 3328 3295

des 3664 3611 3573

apex6 788 774 765

Page 21: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 21

Support the evolution-based heuristic. Provide more sophisticated workflow such

as loop, branching and parallel branches.

Future Work and Conclusion

Page 22: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 22

Thanks for your attention

Page 23: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 23

Question!?

Page 24: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 24

PCB442

Page 25: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 25

PCB442 – 1 branch: Fitness = 55845

Page 26: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 26

PCB442 – 5 branch: Fitness = 54778

Page 27: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 27

PCB442 – 10 branch: Fitness = 54171

Page 28: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 28

ALI535

Page 29: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 29

ALI535 – 1 branch: Fitness = 2136

Page 30: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 30

ALI535 – 5 branch: Fitness = 2140

Page 31: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 31

ALI535 – 10 branch: Fitness = 2250

Page 32: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 32

EIL101

Page 33: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 33

EIL101 – 1 branch: Fitness = 675

Page 34: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 34

EIL101 – 5 branch: Fitness = 679

Page 35: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 35

EIL101 – 10 branch: Fitness = 679

Page 36: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 36

PR1002

Page 37: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 37

PR1002 – 1 branch: Fitness = 293180

Page 38: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 38

PR1002 – 5 branch: Fitness = 288059

Page 39: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 39

PR1002 – 10 branch: Fitness = 288059

Page 40: The Workflow Framework for Metaheuristic  Search  on Grid Environment

HPSC 2009 40

Literal Optimization Result

Optimize

count.blif – IWLS93