Results 1 -
9 of
9
Idwalk : A candidate list strategy with a simple diversification device
- CP 2004: Lecture Notes in Computer Science
, 2004
"... Abstract. This paper presents a new optimization metaheuristic called ID Walk (Intensification/Diversification Walk) that offers advantages for combining simplicity with effectiveness. In addition to the number S of moves, ID Walk uses only one parameter Max which is the maximum number of candidate ..."
Abstract
-
Cited by 9 (4 self)
- Add to MetaCart
(Show Context)
Abstract. This paper presents a new optimization metaheuristic called ID Walk (Intensification/Diversification Walk) that offers advantages for combining simplicity with effectiveness. In addition to the number S of moves, ID Walk uses only one parameter Max which is the maximum number of candidate neighbors studied in every move. This candidate list strategy manages the Max candidates so as to obtain a good tradeoff between intensification and diversification. A procedure has also been designed to tune the parameters automatically. We made experiments on several hard combinatorial optimization problems, and ID Walk compares favorably with correspondingly simple instances of leading metaheuristics, notably tabu search, simulated annealing and Metropolis. Thus, among algorithmic variants that are designed to be easy to program and implement, ID Walk has the potential to become an interesting alternative to such recognized approaches. Our automatic tuning tool has also allowed us to compare several variants of ID Walk and tabu search to analyze which devices (parameters) have the greatest impact on the computation time. A surprising result shows that the specific diversification mechanism embedded in ID Walk is very significant, which motivates examination of additional instances in this new class of “dynamic ” candidate list strategies. 1
Incremental Move for Strip-Packing
- In Proc. ICTAI’07, int. conference on tools with articifcial intelligence, IEEE
, 2007
"... When handling 2D packing problems, numerous incom-plete and complete algorithms maintain a so-called bottom-left (BL) property: every rectangle placed in a container is propped up bottom and left. While it is easy to make a rectangle BL when it is is added in a container, it is more expensive to mai ..."
Abstract
-
Cited by 6 (1 self)
- Add to MetaCart
(Show Context)
When handling 2D packing problems, numerous incom-plete and complete algorithms maintain a so-called bottom-left (BL) property: every rectangle placed in a container is propped up bottom and left. While it is easy to make a rectangle BL when it is is added in a container, it is more expensive to maintain all the placed pieces BL when a rect-angle is removed. This prevents researchers from designing incremental moves for metaheuristics or efficient complete optimization algorithms. This paper investigates the possibility of violating the BL property. Instead, we propose to maintain only the set of “maximal holes”, which allows incremental additions and removals of rectangles. To validate our alternative approach, we have designed an incremental move, maintaining maximal holes, for the strip-packing problem, a variant of the famous 2D bin-packing. We have also implemented a generic metaheuris-tic using this move and standard greedy heuristics. Exper-imental results show that the approach is competitive with the best known incomplete algorithms, especially the other metaheuristics (able to escape from local minima). 1
Strip Packing Based on Local Search and a Randomized Best-Fit
- in "Proc. of First Workshop on Bin Packing and Placement Constraints BPPC’08
, 2008
"... Abstract. We present an incomplete algorithm with no user-defined parameter for handling the strip-packing problem, a variant of the fa-mous 2D bin-packing. The performance of our approach is due to several devices. We propose a move, based on the geometry of the layout, which is made incremental by ..."
Abstract
-
Cited by 5 (0 self)
- Add to MetaCart
(Show Context)
Abstract. We present an incomplete algorithm with no user-defined parameter for handling the strip-packing problem, a variant of the fa-mous 2D bin-packing. The performance of our approach is due to several devices. We propose a move, based on the geometry of the layout, which is made incremental by maintaining the set of maximal holes. For es-caping from local minima, the Intensification Diversification Walk (ID Walk) metaheuristic is used. ID Walk manages only one parameter that is automatically tuned by our tool. We focus here on the greedy heuristics used to perform the moves and to compute the first layout before running the metaheuristic. In partic-ular, we propose a variant of the well-known Best-fit (decreasing) (BF), called RBF, in which the criterion (i.e., height, width, perimeter, surface) changes every time a hole is selected. This simple way to randomize the most efficient greedy strategy is a key for obtaining good bounds while diversifying the layouts.
A Strip Packing Solving Method Using an Incremental Move Based on Maximal Holes
- INTERNATIONAL JOURNAL ON ARTIFICIAL INTELLIGENCE TOOLS
, 2008
"... When handling 2D packing problems, numerous incomplete and complete algorithms maintain a so-called bottom-left (BL) property: no rectangle placed in a container can be moved more left or bottom. While it is easy to make a rectangle BL when it is added in a container, it is more expensive to maintai ..."
Abstract
-
Cited by 3 (0 self)
- Add to MetaCart
When handling 2D packing problems, numerous incomplete and complete algorithms maintain a so-called bottom-left (BL) property: no rectangle placed in a container can be moved more left or bottom. While it is easy to make a rectangle BL when it is added in a container, it is more expensive to maintain all the placed pieces BL when a rectangle is removed. This prevents researchers from designing incremental moves for metaheuristics or efficient complete optimization algorithms. This paper investigates the possibility of violating the BL property. Instead, we propose to maintain the set of maximal holes, which allows incremental additions and removals of rectangles. To validate our alternative approach, we have designed an incremental move, main-taining maximal holes, for the strip packing problem, a variant of the famous 2D bin-packing. We have also implemented a metaheuristic, with no user-defined parameter, using this move and standard greedy heuristics. We have finally designed two variants of this incomplete method. In the first variant, a better first layout is provided by a hyperheuristic proposed by some of the authors. In the second variant, a fast repacking procedure recovering the BL property is occasionally called during the local search.
COMPUTING SOLUTIONS OF THE PAINTSHOP-NECKLACE PROBLEM
, 2011
"... How to assign colors to occurrences of cars in a car factory? How to divide fairly a necklace between thieves who have stolen it? These two questions are addressed in two combinatorial problems that have attracted attention from a theoretical point of view these last years, the first one more by p ..."
Abstract
- Add to MetaCart
(Show Context)
How to assign colors to occurrences of cars in a car factory? How to divide fairly a necklace between thieves who have stolen it? These two questions are addressed in two combinatorial problems that have attracted attention from a theoretical point of view these last years, the first one more by people from the combinatorial optimization community, the second more from the topological combinatorics and computer science point of view. The first problem is the paint shop problem, defined by Epping, Hochstättler and Oertel in 2004. Given a sequence of cars where repetition can occur, and for each car a multiset of colors where the sum of the multiplicities is equal to the number of repetitions of the car in the sequence, decide the color to be applied for each occurrence of each car so that each color occurs with the multiplicity that has been assigned. The goal is to minimize the number of color changes in the sequence. The second problem, highly related to the first one, takes its origin in a famous theorem found by Alon in 1987 stating that a necklace with t types of beads and qau occurrences of each type u (au is a positive integer) can always be fairly split between q thieves with at most t(q − 1) cuts. An intriguing aspect of this theorem lies in the fact that its classical proof is completely nonconstructive. Designing an algorithm that computes theses cuts is not an easy task, and remains mostly open. The main purpose of the present paper is to make a step in a more operational direction for these two problems by discussing practical ways to compute solutions for instances of various sizes. Moreover, it starts with an exhaustive survey on the algorithmic aspects of them, and some new results are proved.
Database and Artificial Intelligence Group (184/2)
, 2006
"... In the area of computer science heuristics are very important when it comes to solving NP hard problems. Though they neither provide an accurate method of finding a solution nor a provable approximation like Approximation algorithms, they nevertheless find reasonable good solutions to NP hard proble ..."
Abstract
- Add to MetaCart
(Show Context)
In the area of computer science heuristics are very important when it comes to solving NP hard problems. Though they neither provide an accurate method of finding a solution nor a provable approximation like Approximation algorithms, they nevertheless find reasonable good solutions to NP hard problems usually while consuming a reasonable amount of computation time and memory. A well known class of Heuristic methods are Local Search techniques. Those techniques focus on starting with a solution, creating a neighbourhood for the solution by applying small changes to the solution, evaluating the solutions of the neighbourhood and choosing a solution of the neighbourhood for the next iteration. The algorithm terminates when the termination condition is fulfilled. Whenever a problem is solved with Tabu Search in a program the structural and algorithmic definitions of Tabu Search have to be implemented. A solution to avoid reimplementing Tabu Search again and again, is to create a framework providing a basic design for Tabu Search. To remove the cumbersome process of
