Shortcomings of Traditional Backtrack Search on Large, Tight CSPs: A Real-world Example Venkata Praveen Guddeti and Berthe Y. Choueiry The combination.

Slides:



Advertisements
Similar presentations
Heuristic Search techniques
Advertisements

1 Restart search techniques for Employee Timetabling Problems Amnon Meisels And Eliezer Kaplansky Ben-Gurion University.
An Interactive System for Hiring & Managing Graduate Teaching Assistants Ryan Lim Venkata Praveen Guddeti Berthe Y. Choueiry Constraint Systems Laboratory.
Best-First Search: Agendas
Constraint Processing Techniques for Improving Join Computation: A Proof of Concept Anagh Lal & Berthe Y. Choueiry Constraint Systems Laboratory Department.
Foundations of Constraint Processing, Spring 2008 Evaluation to BT SearchApril 16, Foundations of Constraint Processing CSCE421/821, Spring 2008:
Constraint Systems Laboratory Oct 21, 2004Guddeti: MS thesis defense1 An Improved Restart Strategy for Randomized Backtrack Search Venkata P. Guddeti Constraint.
CP Formal Models of Heavy-Tailed Behavior in Combinatorial Search Hubie Chen, Carla P. Gomes, and Bart Selman
Foundations of Constraint Processing, Fall 2004 November 8, 2004Ordering heuristics1 Foundations of Constraint Processing CSCE421/821, Fall 2004:
An Approximation of Generalized Arc-Consistency for Temporal CSPs Lin Xu and Berthe Y. Choueiry Constraint Systems Laboratory Department of Computer Science.
Iterative Improvement Techniques for Solving Tight Constraint Satisfaction Problems Hui Zou Constraint Systems Laboratory Department of Computer Science.
Heavy-Tailed Behavior and Search Algorithms for SAT Tang Yi Based on [1][2][3]
Improving Backtrack Search For Solving the TCSP Lin Xu and Berthe Y. Choueiry Constraint Systems Laboratory Department of Computer Science and Engineering.
CPSC 322, Lecture 12Slide 1 CSPs: Search and Arc Consistency Computer Science cpsc322, Lecture 12 (Textbook Chpt ) January, 29, 2010.
Shirokuro : A Backtracking Approach Benjamin Bush Faculty Advisors: Dr. Russ Abbott, Dr. Gary Brookfield Department of Computer Science, Department of.
A Constraint Satisfaction Problem (CSP) is a combinatorial decision problem defined by a set of variables, a set of domain values for these variables,
 i may require adding new constraints, except for… o i =1  domain filtering o i =   constraint filtering Robert Woodward & Berthe Y. Choueiry Constraint.
Multi-Agent Based Search vs. Local Search and Backtrack Search for Solving Tight CSPs: A Practical Case Study Hui Zou and Berthe Y. Choueiry Constraint.
Foundations of Constraint Processing Evaluation to BT Search 1 Foundations of Constraint Processing CSCE421/821, Spring
Ryan Kinworthy 2/26/20031 Chapter 7- Local Search part 1 Ryan Kinworthy CSCE Advanced Constraint Processing.
An Interactive, Constraint-Based System for Task Allocation in an Academic Environment Ryan Lim, Praveen Guddeti, and Berthe Y. Choueiry Constraint Systems.
Solvable problem Deviation from best known solution [%] Percentage of test runs ERA RDGR RGR LS Over-constrained.
Efficient Techniques for Searching the Temporal CSP Lin Xu and Berthe Y. Choueiry Constraint Systems Laboratory Department of Computer Science and Engineering.
Foundations of Constraint Processing, Fall 2005 Sep 20, 2005BT: A Theoretical Evaluation1 Foundations of Constraint Processing CSCE421/821, Fall 2005:
A Constraint Satisfaction Problem (CSP) is a combinatorial decision problem defined by a set of variables, a set of domain values for these variables,
Support: UCARE grant awarded to Chris Reeson & CAREER Award # from the National Science Foundation. To the public: –Illustrate the power of CP For.
Constraint Systems Laboratory March 26, 2007Reeson–Undergraduate Thesis1 Using Constraint Processing to Model, Solve, and Support Interactive Solving of.
1 Bandwidth Allocation Planning in Communication Networks Christian Frei & Boi Faltings Globecom 1999 Ashok Janardhanan.
Foundations of Constraint Processing, Fall 2004 November 18, 2004More on BT search1 Foundations of Constraint Processing CSCE421/821, Fall 2004:
Hardness-Aware Restart Policies Yongshao Ruan, Eric Horvitz, & Henry Kautz IJCAI 2003 Workshop on Stochastic Search.
Distributions of Randomized Backtrack Search Key Properties: I Erratic behavior of mean II Distributions have “heavy tails”.
1 Problem Solving and Searching CS 171/271 (Chapter 3) Some text and images in these slides were drawn from Russel & Norvig’s published material.
SDQ for data set 5 (500 runs, 5 mins each) PDF for relative solution size for data set 5 (500 runs, 5 mins each) Statistics of solution size (500 runs,
Motivation & Goal SAT and Constraint Processing (CP) are fundamental areas of Computer Science that address the same computational questions. Compare SAT.
Department of Mathematical Sciences February 9, /11 Art Duval and Helmut Knaust Student Outcomes in Mathematics.
Problem Solving with Constraints Lookahead Schemas 1 Foundations of Constraint Processing CSCE496/896, Fall
Constraint Systems Laboratory 11/26/2015Zhang: MS Project Defense1 OPRAM: An Online System for Assigning Capstone Course Students to Sponsored Projects.
Constraint Systems Laboratory Presented by: Robert J. Woodward, Amanda Swearngin 1 Berthe Y. Choueiry 2 Eugene C. Freuder 3 1 ESQuaReD Laboratory, University.
Quality of LP-based Approximations for Highly Combinatorial Problems Lucian Leahu and Carla Gomes Computer Science Department Cornell University.
1 Attractive Mathematical Representations Of Decision Problems Warren Adams 11/04/03.
Foundations of Constraint Processing, Fall 2004 October 3, 2004Interchangeability in CSPs1 Foundations of Constraint Processing CSCE421/821, Fall 2004:
Balance and Filtering in Structured Satisfiability Problems Henry Kautz University of Washington joint work with Yongshao Ruan (UW), Dimitris Achlioptas.
Robust Planning using Constraint Satisfaction Techniques Daniel Buettner and Berthe Y. Choueiry Constraint Systems Laboratory Department of Computer Science.
At the beginning of each semester, CSE hires a number of Graduate Teaching Assistants (GTAs) as graders, lab supervisors, and instructors. The department.
Maintaining Consistency in Resource Allocation Trevor Janke & Berthe Y. Choueiry Goals  Study & refine the constraint model of the existing Graduate TAs.
1 CSC 384 Lecture Slides (c) , C. Boutilier and P. Poupart CSC384: Lecture 16  Last time Searching a Graphplan for a plan, and relaxed plan heuristics.
ERA on an over-constrained problem A Constraint-Based System for Hiring & Managing Graduate Teaching Assistants Ryan Lim, Praveen Venkata Guddeti, and.
Wolfgang Runte Slide University of Osnabrueck, Software Engineering Research Group Wolfgang Runte Software Engineering Research Group Institute.
Consistency Methods for Temporal Reasoning
Robert Glaubius and Berthe Y. Choueiry
Empirical Comparison of Preprocessing and Lookahead Techniques for Binary Constraint Satisfaction Problems Zheying Jane Yang & Berthe Y. Choueiry Constraint.
Extensions to BT Search Foundations of Constraint Processing
Extensions to BT Search Foundations of Constraint Processing
Intelligent Backtracking Algorithms: A Theoretical Evaluation
Evaluation of (Deterministic) BT Search Algorithms
Extensions to BT Search Foundations of Constraint Processing
Extensions to BT Search Foundations of Constraint Processing
Chapter 5: General search strategies: Look-ahead
Evaluation of (Deterministic) BT Search Algorithms
Intelligent Backtracking Algorithms: A Theoretical Evaluation
Intelligent Backtracking Algorithms: A Theoretical Evaluation
Constraint Satisfaction Problems & Its Application in Databases
Intelligent Backtracking Algorithms: A Theoretical Evaluation
Intelligent Backtracking Algorithms: A Theoretical Evaluation
Artificial Intelligence
Evaluation of (Deterministic) BT Search Algorithms
Extensions to BT Search Foundations of Constraint Processing
Intelligent Backtracking Algorithms: A Theoretical Evaluation
A Qualitative Analysis of Search Behavior: A Visual Approach
Area Coverage Problem Optimization by (local) Search
Presentation transcript:

Shortcomings of Traditional Backtrack Search on Large, Tight CSPs: A Real-world Example Venkata Praveen Guddeti and Berthe Y. Choueiry The combination of the 4 variable ordering heuristics and the 3 value ordering heuristic yielded 12 ordering heuristics. Three real-world data sets of spring 2001, fall2001 and fall2002 were used for the experiments. Thus, 48 experiments were carried out. Each experiment was run for 10 minutes. Results were noted for both the best solution found and for full 10 minute run. Constraint Systems Laboratory Department of Computer Science and Engineering, University of Nebraska-Lincoln Abstract Problems of systematic BT search Systematic Search Current investigations Methodology References Experiments Systematic Search: Forward checking algorithm augmented with various ordering heuristics. We compared the ordering heuristics according to five criteria: 1.Unassigned courses: the number of courses that are not assigned a GTA. 2.CC: the number of constraint checks. 3.NNV: the number of nodes visited. 4.Number of backtracks: the number of backtracks done. Unassigned variables: The table below shows the number of unassigned variables and the CPU run time taken for the solution. We are testing various restart strategies based on Fixed restarts and Dynamic restarts [2]. Limited BT search: Keeping a count of the amount of backtracking done so far and abandoning the search if more than the cutoff value [3]. Credit based search: Each search path is assigned some credit. This credit may or may not be uniform. If credit is over than perform deterministic search and/or small amount of local search [3]. 1.R. Glaubius and B. Y. Choueiry, “Constraint Modeling and Reformulation in the Context of Academic Task Assignment”. In Working Notes of the Workshop Modeling and Solving Problems with Constraints, ECAI H. Kautz, E. Horvitz, Y. Ruan, C. Gomes, B. Selman; “Dynamic Restart Policies”. In Proceedings of the Eighteenth National Conference on Artificial Intelligence, Edmonton, Alberta, July AAAI Press. 3.H. Simonis, Invited Talk at the Constraint Systems Laboratory, UNL, November The Graduate Teaching Assistant (GTA) assignment project [1] aims at matching GTAs to academic tasks according to their abilities and availability. It is, is an on-going applied research project conducted at the Constraint Systems Laboratory as a service to the Department of Computer Science and Engineering and as a hands-on tool for training undergraduate and graduate students in advanced Constraint Processing techniques. In practice, the problem is always tight and often over-constrained, in that there are more courses to serve than there are GTAs hired to help. Also there are many non-convex constraints that restrict allowable assignments and combinations of GTAs to classes. Finally. we have enriched the application by allowing GTAs to express their preferences on a scale from 0 (unable to handle) to 5 (highest preference). Our goal is to explore and compare the performance of various backtrack search techniques: systematic search [1] and random backtrack search [2]. Variable Ordering Heuristics: 2 variable selection heuristics were implemented using least domain and domain-degree ratio as selection criteria. These heuristics were ordered statically and dynamically. Thus, we had 4 variable ordering heuristics: 1.Static Least Domain ( SLD ) 2.Static Domain Degree Ratio ( SDD ) 3.Dynamic Least Domain ( DLD ) 4.Dynamic Domain Degree Ratio ( DDD ) Thrashing of BT search: Value Ordering Heuristics: 3 value ordering heuristics: 1.First in last out ( FIL ): The first available value is selected. 2.Highest preference value ( PREFERENCE ): The value having the highest preference is selected. 3.Least occurring value ( OCCURRENCE ): The value occurring the least number of times in the future variables is selected. Variables in the spring 2001 data set: 69 Average shallowest level of BT: 48 Thrashing in 30% of the search tree. Many ordering heuristics have also been implemented to improve the performance of search and the embedded forward checking (FC) mechanism implemented for the GTA problem [1]. We experimented with various data sets of the GTA assignment problem to analyze and compare the performance of the various ordering heuristics implemented. We report our observations and summarize our analysis of the shortcomings of traditional backtrack search methods on large over-constrained CSPs. We show that the performance of value and variable ordering heuristics depend on the problem instance. And the BT mechanism gets “thrashed” in a small portion of the search space; unable to undo early choice, regardless of the ordering heuristic used. Random Ordering Heuristics: 2 heuristics using randomness: 1.All Random: The value and variable selection were done randomly. 2.Hybrid heuristic: Random heuristics in combination with the above mentioned ordering heuristics. Problems of random BT search The search tree of the GTA problem is so large that even random BT search is not enough to avoid the thrashing during search. Although the solutions vary with each run, the shallowest BT level does not improve.