An Evolutionary Hyper-heuristic for the Software Project Scheduling Problem

Wu, Xiuli; Consoli, Pietro A.; Minku, Leandro L.; Ochoa, Gabriela; Yao, Xin

doi:10.1007/978-3-319-45823-6_4

Cited by 21 publications

(14 citation statements)

References 16 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This typically involves the training of machine learning models on performance data of algorithms in combination with instances given as feature data. In software engineering, this has been recently used as an approach for the Software Project Scheduling Problem [22], [23]. The field of per-instance configuration has received much attention recently, and we refer the interested reader to a recent updated survey article [24].…”

Section: Per-corpus Configurationmentioning

confidence: 99%

Predicting Good Configurations for GitHub and Stack Overflow Topic Models

Wagner¹

2019

2019 IEEE/ACM 16th International Conference on Mining Software Repositories (MSR)

View full text Add to dashboard Cite

Software repositories contain large amounts of textual data, ranging from source code comments and issue descriptions to questions, answers, and comments on Stack Overflow. To make sense of this textual data, topic modelling is frequently used as a text-mining tool for the discovery of hidden semantic structures in text bodies. Latent Dirichlet allocation (LDA) is a commonly used topic model that aims to explain the structure of a corpus by grouping texts. LDA requires multiple parameters to work well, and there are only rough and sometimes conflicting guidelines available on how these parameters should be set. In this paper, we contribute (i) a broad study of parameters to arrive at good local optima for GitHub and Stack Overflow text corpora, (ii) an a-posteriori characterisation of text corpora related to eight programming languages, and (iii) an analysis of corpus feature importance via per-corpus LDA configuration. We find that (1) popular rules of thumb for topic modelling parameter configuration are not applicable to the corpora used in our experiments, (2) corpora sampled from GitHub and Stack Overflow have different characteristics and require different configurations to achieve good model fit, and (3) we can predict good configurations for unseen corpora reliably. These findings support researchers and practitioners in efficiently determining suitable configurations for topic modelling when analysing textual data contained in software repositories.

show abstract

Section: Per-corpus Configurationmentioning

confidence: 99%

Predicting Good Configurations for GitHub and Stack Overflow Topic Models

Wagner¹

2019

2019 IEEE/ACM 16th International Conference on Mining Software Repositories (MSR)

View full text Add to dashboard Cite

show abstract

“…This typically involves the training of machine learning models on performance data of algorithms in combination with instances given as feature data. In software engineering, this has been recently used as for the Software Project Scheduling Problem [92,102]. The field of per-instance configuration has received much attention recently, and we refer the interested reader to a recent updated survey article [50].…”

Section: Research Directionmentioning

confidence: 99%

Better software analytics via “DUO”: Data mining algorithms using/used-by optimizers

et al. 2020

View full text Add to dashboard Cite

This paper claims that a new field of empirical software engineering research and practice is emerging: data mining using/used-by optimizers for empirical studies, or DUO. For example, data miners can generate the models that are explored by optimizers. Also, optimizers can advise how to best adjust the control parameters of a data miner. This combined approach acts like an agent leaning over the shoulder of an analyst that advises "ask this question next" or "ignore that problem, it is not relevant to your goals". Further, those agents can help us build "better" predictive models, where "better" can be either greater predictive accuracy, or faster modeling time (which, in turn, enables the exploration of a wider range of options). We also caution that the era of papers that just use data miners is coming to an end. Results obtained from an unoptimized data miner can be quickly refuted, just by applying an optimizer to produce a different (and better performing) model. Our conclusion, hence, is that for software analytics it is possible, useful and necessary to combine data mining and optimization using DUO.

show abstract

“…By analysing the source of feedback information, three modules can be considered: online, offline, and no-learning. Choice function [5,47], reinforcement learning [48], TS [2,6,[49][50][51][52], and FRR-MAB [2,[53][54][55][56][57] are examples for online selection strategies, and simple random, random descent, RP, etc., are viewed as no-learning methods. Several metaheuristic-based strategies for designing hyperheuristics have been proposed in the literature: ant colony-, particle-, and quantum-inspired hyperheuristic.…”

Section: Hyperheuristicmentioning

confidence: 99%

“…With the popularity of hyperheuristic, it has been widely applied in practice, such as 2D regular and irregular packing problems [73], nurse rostering [74], vehicle routing problem [54], construction levelling problem [52], software project scheduling problem [56], t-ways test suite generation [6,75], deriving products for variability test of feature models [55], fast machine reassignment [76], and timetabling [77][78][79][80][81]. We refer interested readers to these papers [43,82,83] for extensive review on hyperheuristic.…”

Section: Hyperheuristicmentioning

confidence: 99%

Evolutionary Hyperheuristics for Location-Routing Problem with Simultaneous Pickup and Delivery

et al. 2020

View full text Add to dashboard Cite

This paper presents an evolution-based hyperheuristic (EHH) for addressing the capacitated location-routing problem (CLRP) and one of its more practicable variants, namely, CLRP with simultaneous pickup and delivery (CLRPSPD), which are significant and NP-hard model in the complex logistics system. The proposed approaches manage a pool of low-level heuristics (LLH), implementing a set of simple, cheap, and knowledge-poor operators such as “shift” and “swap” to guide the search. Quantum (QS), ant (AS), and particle-inspired (PS) high-level learning strategies (HLH) are developed as evolutionary selection strategies (ESs) to improve the performance of the hyperheuristic framework. Meanwhile, random permutation (RP), tabu search (TS), and fitness rate rank-based multiarmed bandit (FRR-MAB) are also introduced as baselines for comparisons. We evaluated pairings of nine different selection strategies and four acceptance mechanisms and monitored the performance of the first four outstanding pairs in 36 pairs by solving three sets of benchmark instances from the literature. Experimental results show that the proposed approaches outperform most fine-tuned bespoke state-of-the-art approaches in the literature, and PS-AM and AS-AM perform better when compared to the rest of the pairs in terms of obtaining a good trade-off of solution quality and computing time.

show abstract

An Evolutionary Hyper-heuristic for the Software Project Scheduling Problem

Cited by 21 publications

References 16 publications

Predicting Good Configurations for GitHub and Stack Overflow Topic Models

Predicting Good Configurations for GitHub and Stack Overflow Topic Models

Better software analytics via “DUO”: Data mining algorithms using/used-by optimizers

Evolutionary Hyperheuristics for Location-Routing Problem with Simultaneous Pickup and Delivery

Contact Info

Product

Resources

About