Pub Date : 2020-06-12DOI: 10.4230/LIPIcs.CP.2021.33
Chaitanya K. Joshi, Quentin Cappart, Louis-Martin Rousseau, T. Laurent
End-to-end training of neural network solvers for graph combinatorial optimization problems such as the Travelling Salesperson Problem (TSP) have seen a surge of interest recently, but remain intractable and inefficient beyond graphs with few hundreds of nodes. While state-of-the-art learning-driven approaches for TSP perform closely to classical solvers when trained on trivially small sizes, they are unable to generalize the learnt policy to larger instances at practical scales. This work presents an end-to-end neural combinatorial optimization pipeline that unifies several recent papers in order to identify the inductive biases, model architectures and learning algorithms that promote generalization to instances larger than those seen in training. Our controlled experiments provide the first principled investigation into such zero-shot generalization, revealing that extrapolating beyond training data requires rethinking the neural combinatorial optimization pipeline, from network layers and learning paradigms to evaluation protocols. Additionally, we analyze recent advances in deep learning for routing problems through the lens of our pipeline and provide new directions to stimulate future research.
{"title":"Learning the travelling salesperson problem requires rethinking generalization","authors":"Chaitanya K. Joshi, Quentin Cappart, Louis-Martin Rousseau, T. Laurent","doi":"10.4230/LIPIcs.CP.2021.33","DOIUrl":"https://doi.org/10.4230/LIPIcs.CP.2021.33","url":null,"abstract":"End-to-end training of neural network solvers for graph combinatorial optimization problems such as the Travelling Salesperson Problem (TSP) have seen a surge of interest recently, but remain intractable and inefficient beyond graphs with few hundreds of nodes. While state-of-the-art learning-driven approaches for TSP perform closely to classical solvers when trained on trivially small sizes, they are unable to generalize the learnt policy to larger instances at practical scales. This work presents an end-to-end neural combinatorial optimization pipeline that unifies several recent papers in order to identify the inductive biases, model architectures and learning algorithms that promote generalization to instances larger than those seen in training. Our controlled experiments provide the first principled investigation into such zero-shot generalization, revealing that extrapolating beyond training data requires rethinking the neural combinatorial optimization pipeline, from network layers and learning paradigms to evaluation protocols. Additionally, we analyze recent advances in deep learning for routing problems through the lens of our pipeline and provide new directions to stimulate future research.","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"27 1","pages":"70 - 98"},"PeriodicalIF":1.6,"publicationDate":"2020-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42327443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we present a summary of XCSP3, together with its ecosystem. XCSP3 is a format used to build integrated representations of combinatorial constrained problems. Interestingly, XCSP3 preserves the structure of models, by handling arrays of variables and groups/blocks of constraints, which makes it rather unique in the literature. Furthermore, the ecosystem of XCSP3 is well supplied: it includes companion tools (parsers and checkers), a website with a search engine for selecting and downloading instances, and competitions of solvers. The Java-based modeling API, called JvCSP3, is the last developed piece of this complete production chain.
{"title":"XCSP 3 and its ecosystem","authors":"Gilles Audemard, Frédéric Boussemart, Christophe Lecoutre, Cédric Piette, Olivier Roussel","doi":"10.1007/s10601-019-09307-9","DOIUrl":"https://doi.org/10.1007/s10601-019-09307-9","url":null,"abstract":"In this paper, we present a summary of XCSP<sup>3</sup>, together with its ecosystem. XCSP<sup>3</sup> is a format used to build integrated representations of combinatorial constrained problems. Interestingly, XCSP<sup>3</sup> preserves the structure of models, by handling arrays of variables and groups/blocks of constraints, which makes it rather unique in the literature. Furthermore, the ecosystem of XCSP<sup>3</sup> is well supplied: it includes companion tools (parsers and checkers), a website with a search engine for selecting and downloading instances, and competitions of solvers. The Java-based modeling API, called JvCSP<sup>3</sup>, is the last developed piece of this complete production chain.","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"27 3","pages":""},"PeriodicalIF":1.6,"publicationDate":"2020-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138503077","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-15DOI: 10.1007/s10601-019-09306-w
Jaime E. González, A. Ciré, Andrea Lodi, Louis-Martin Rousseau
{"title":"Integrated integer programming and decision diagram search tree with an application to the maximum independent set problem","authors":"Jaime E. González, A. Ciré, Andrea Lodi, Louis-Martin Rousseau","doi":"10.1007/s10601-019-09306-w","DOIUrl":"https://doi.org/10.1007/s10601-019-09306-w","url":null,"abstract":"","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"25 1","pages":"23 - 46"},"PeriodicalIF":1.6,"publicationDate":"2020-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/s10601-019-09306-w","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"52195928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-20DOI: 10.1007/s10601-019-09305-x
Mohamed Amine Omrani, Wady Naanaa
Although graphs are widely used to encode and solve various computational problems, little research exists on constrained graph construction. The current research was carried out to shed light on the problem of generating graphs, where the construction process is guided by various structural restrictions, like vertex degrees, proximity among vertices, and imposed and forbidden patterns. The main contribution of this paper is an encoding of the constrained graph generation problem in terms of a constraint satisfaction problem (CSP). This approach is motivated by the flurry of efficient solution algorithms available within the constraint programming (CP) framework. The obtained encoding has given rise to the CP-MolGen program, a new open source program dedicated to the generation of molecular graphs with imposed and forbidden fragments. Experimental results on several real-world molecular graph generation instances have shown the effectiveness and efficiency of the proposed program, especially the benefits of forbidding cyclic patterns as induced subgraphs.
{"title":"Constraints for generating graphs with imposed and forbidden patterns: an application to molecular graphs","authors":"Mohamed Amine Omrani, Wady Naanaa","doi":"10.1007/s10601-019-09305-x","DOIUrl":"https://doi.org/10.1007/s10601-019-09305-x","url":null,"abstract":"Although graphs are widely used to encode and solve various computational problems, little research exists on constrained graph construction. The current research was carried out to shed light on the problem of generating graphs, where the construction process is guided by various structural restrictions, like vertex degrees, proximity among vertices, and imposed and forbidden patterns. The main contribution of this paper is an encoding of the constrained graph generation problem in terms of a constraint satisfaction problem (CSP). This approach is motivated by the flurry of efficient solution algorithms available within the constraint programming (CP) framework. The obtained encoding has given rise to the CP-MolGen program, a new open source program dedicated to the generation of molecular graphs with imposed and forbidden fragments. Experimental results on several real-world molecular graph generation instances have shown the effectiveness and efficiency of the proposed program, especially the benefits of forbidding cyclic patterns as induced subgraphs.","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"28 1","pages":"1 - 22"},"PeriodicalIF":1.6,"publicationDate":"2019-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138503076","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-05DOI: 10.1007/s10601-020-09317-y
Ondřej Benedikt, I. Módos, Z. Hanzálek
{"title":"Power of pre-processing: production scheduling with variable energy pricing and power-saving states","authors":"Ondřej Benedikt, I. Módos, Z. Hanzálek","doi":"10.1007/s10601-020-09317-y","DOIUrl":"https://doi.org/10.1007/s10601-020-09317-y","url":null,"abstract":"","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"25 1","pages":"300 - 318"},"PeriodicalIF":1.6,"publicationDate":"2019-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/s10601-020-09317-y","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46428023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Correction to: The potential of quantum annealing for rapid solution structure identification","authors":"Yuchen Pang, Carleton Coffrin, A. Lokhov, Marc Vuffray","doi":"10.2172/1599019","DOIUrl":"https://doi.org/10.2172/1599019","url":null,"abstract":"A Correction to this paper has been published: https://doi.org/10.1007/s10601-021-09320-x","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"26 1","pages":"107"},"PeriodicalIF":1.6,"publicationDate":"2019-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41649612","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-20DOI: 10.1007/s10601-019-09304-y
V. Levit, Zohar Komarovsky, Tal Grinshpoun, A. Bazzan, A. Meisels
{"title":"Incentive-based search for equilibria in boolean games","authors":"V. Levit, Zohar Komarovsky, Tal Grinshpoun, A. Bazzan, A. Meisels","doi":"10.1007/s10601-019-09304-y","DOIUrl":"https://doi.org/10.1007/s10601-019-09304-y","url":null,"abstract":"","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"24 1","pages":"288 - 319"},"PeriodicalIF":1.6,"publicationDate":"2019-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/s10601-019-09304-y","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"52195904","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-04-29DOI: 10.1007/s10601-019-09303-z
A. Palmieri, Arnaud Lallouet, Luc Pons
{"title":"Constraint Games for stable and optimal allocation of demands in SDN","authors":"A. Palmieri, Arnaud Lallouet, Luc Pons","doi":"10.1007/s10601-019-09303-z","DOIUrl":"https://doi.org/10.1007/s10601-019-09303-z","url":null,"abstract":"","PeriodicalId":55211,"journal":{"name":"Constraints","volume":"24 1","pages":"252 - 287"},"PeriodicalIF":1.6,"publicationDate":"2019-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/s10601-019-09303-z","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46505728","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：