Generation of tests for programming challenge tasks using multi-objective optimization

Maxim Buzdalov; Arina Buzdalova; Irina Petrova

doi:10.1145/2464576.2482746

Generation of tests for programming challenge tasks using multi-objective optimization

Maxim Buzdalov, Arina Buzdalova, Irina Petrova

ITMO University

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (Non-Journal item)

13 Citations (Scopus)

Abstract

In this paper, an evolutionary approach to generation of test cases for programming challenge tasks is investigated. Multi-objective and single-objective evolutionary algorithms, as well as helper-objective selection strategies, are compared. Particularly, a previously proposed method of choosing between helper-objectives with reinforcement learning is considered. This method is applied to the multi-objective evolutionary algorithm for the first time. Results of the experiment show that the most reasonable approach for the considered problem is using multi-objective evolutionary algorithm with automated helper-objective selection.

Original language	English
Title of host publication	GECCO '13 Companion
Subtitle of host publication	Proceedings of the 15th annual conference companion on Genetic and evolutionary computation
Editors	Christian Blum
Publisher	Association for Computing Machinery
Pages	1655-1658
Number of pages	4
ISBN (Print)	978-1-4503-1964-5
DOIs	https://doi.org/10.1145/2464576.2482746
Publication status	Published - 06 Jul 2013
Externally published	Yes
Event	GECCO 2013 - Genetic and Evolutionary Computation Conference - Amsterdam, Netherlands Duration: 06 Jul 2013 → 10 Jul 2013

Conference

Conference	GECCO 2013 - Genetic and Evolutionary Computation Conference
Country/Territory	Netherlands
City	Amsterdam
Period	06 Jul 2013 → 10 Jul 2013

Keywords

genetic algorithms
helper-objectives
multi-objective
programming challenges
reinforcement learning
testing

Access to Document

10.1145/2464576.2482746

Permanent link

Persistent link

Cite this

@inproceedings{d6688303b7e648a68812c06f17c4a22b,

title = "Generation of tests for programming challenge tasks using multi-objective optimization",

abstract = "In this paper, an evolutionary approach to generation of test cases for programming challenge tasks is investigated. Multi-objective and single-objective evolutionary algorithms, as well as helper-objective selection strategies, are compared. Particularly, a previously proposed method of choosing between helper-objectives with reinforcement learning is considered. This method is applied to the multi-objective evolutionary algorithm for the first time. Results of the experiment show that the most reasonable approach for the considered problem is using multi-objective evolutionary algorithm with automated helper-objective selection.",

keywords = "genetic algorithms, helper-objectives, multi-objective, programming challenges, reinforcement learning, testing",

author = "Maxim Buzdalov and Arina Buzdalova and Irina Petrova",

year = "2013",

month = jul,

day = "6",

doi = "10.1145/2464576.2482746",

language = "English",

isbn = "978-1-4503-1964-5",

pages = "1655--1658",

editor = "Christian Blum",

booktitle = "GECCO '13 Companion",

publisher = "Association for Computing Machinery",

address = "United States of America",

note = "GECCO 2013 - Genetic and Evolutionary Computation Conference ; Conference date: 06-07-2013 Through 10-07-2013",

}

Buzdalov, M , Buzdalova, A & Petrova, I 2013, Generation of tests for programming challenge tasks using multi-objective optimization. in C Blum (ed.), GECCO '13 Companion: Proceedings of the 15th annual conference companion on Genetic and evolutionary computation. Association for Computing Machinery, pp. 1655-1658, GECCO 2013 - Genetic and Evolutionary Computation Conference, Amsterdam, Netherlands, 06 Jul 2013. https://doi.org/10.1145/2464576.2482746

Generation of tests for programming challenge tasks using multi-objective optimization. / Buzdalov, Maxim ; Buzdalova, Arina; Petrova, Irina.
GECCO '13 Companion: Proceedings of the 15th annual conference companion on Genetic and evolutionary computation. ed. / Christian Blum. Association for Computing Machinery, 2013. p. 1655-1658.

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (Non-Journal item)

TY - GEN

T1 - Generation of tests for programming challenge tasks using multi-objective optimization

AU - Buzdalov, Maxim

AU - Buzdalova, Arina

AU - Petrova, Irina

PY - 2013/7/6

Y1 - 2013/7/6

N2 - In this paper, an evolutionary approach to generation of test cases for programming challenge tasks is investigated. Multi-objective and single-objective evolutionary algorithms, as well as helper-objective selection strategies, are compared. Particularly, a previously proposed method of choosing between helper-objectives with reinforcement learning is considered. This method is applied to the multi-objective evolutionary algorithm for the first time. Results of the experiment show that the most reasonable approach for the considered problem is using multi-objective evolutionary algorithm with automated helper-objective selection.

AB - In this paper, an evolutionary approach to generation of test cases for programming challenge tasks is investigated. Multi-objective and single-objective evolutionary algorithms, as well as helper-objective selection strategies, are compared. Particularly, a previously proposed method of choosing between helper-objectives with reinforcement learning is considered. This method is applied to the multi-objective evolutionary algorithm for the first time. Results of the experiment show that the most reasonable approach for the considered problem is using multi-objective evolutionary algorithm with automated helper-objective selection.

KW - genetic algorithms

KW - helper-objectives

KW - multi-objective

KW - programming challenges

KW - reinforcement learning

KW - testing

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-84882419302&origin=resultslist

U2 - 10.1145/2464576.2482746

DO - 10.1145/2464576.2482746

M3 - Conference Proceeding (Non-Journal item)

SN - 978-1-4503-1964-5

SP - 1655

EP - 1658

BT - GECCO '13 Companion

A2 - Blum, Christian

PB - Association for Computing Machinery

T2 - GECCO 2013 - Genetic and Evolutionary Computation Conference

Y2 - 6 July 2013 through 10 July 2013

ER -

Generation of tests for programming challenge tasks using multi-objective optimization

Abstract

Conference

Keywords

Access to Document

Permanent link

Other files and links

Fingerprint

Cite this