A comprehensive database of Nature-Inspired Algorithms

These data contain a comprehensive collection of all Nature-Inspired Algorithms. This collection is a result of two corresponding surveys, where all Nature-Inspired Algorithms that have been published to-date were gathered and preliminary data acquired. The rapidly increasing number of nature-inspired approaches makes it hard for interested researchers to keep up. Moreover, a proper taxonomy is necessary, based on specific features of the algorithms. Different taxonomies and useful insight into the application areas that the algorithms have coped with is given through these data. This article provides a detailed description of the above mentioned collection.


a b s t r a c t
These data contain a comprehensive collection of all Nature-Inspired Algorithms. This collection is a result of two corresponding surveys, where all Nature-Inspired Algorithms that have been published to-date were gathered and preliminary data acquired. The rapidly increasing number of natureinspired approaches makes it hard for interested researchers to keep up. Moreover, a proper taxonomy is necessary, based on specific features of the algorithms. Different taxonomies and useful insight into the application areas that the algorithms have coped with is given through these data. This article provides a detailed description of the above mentioned collection.
© 2020 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license.
( http://creativecommons.org/licenses/by/4.0/ ) Table   Subject Artificial Intelligence Specific subject area Nature-Inspired Algorithms Type of data csv file How data were acquired Data were acquired through research in documents and records from International Journals and Conferences.

Data format Raw: csv file Parameters for data collection
Only Nature-Inspired Algorithms are included in this dataset, based on the definition given by [1] : "The term nature refers to any part of the physical universe which is not a product of intentional human design". To select the algorithms meeting the above definition properly, the authors read the initial study proposing the algorithm and excluded methods inspired by social theory (Political Optimiser, etc.), sports (i.e. the League Championship Algorithm) or the result of intentional human design (such as the Fireworks Algorithm).

Description of data collection
These data were collected through web research.

Value of the Data
• These data consist of the first comprehensive list of Nature-Inspired Algorithms, where the main information for each algorithm can be found (year, authors, Journal or Conference where it was initially presented, applications that were tackled in the initial work, etc.). Moreover, information is included on the application areas that each algorithm has been applied to. • Interested audiences can benefit from this data set, while also, researchers who are interested in narrowing down their choices when trying to find a proper algorithm for their application. Furthermore, the algorithms included in this database will benefit and be introduced to a greater number of readers. • Useful insights can be extracted from these data. Based on this data set, more secondary data could be carried out that will lead to adequate survey studies. • Furthermore, the field of Nature-Inspired Intelligence would benefit from this data set. New hybrid schemes could be developed based on the provided information of the data, while also further research can be done on the features that an algorithm should have to cope with a specific problem or problem area. • Finally, the provided data set could even be used as a benchmark for future surveys that focus on a specific application area. Additionally, these data also allow the citation and bibliometric analysis of papers in the area of Nature-Inspired Computing.

Data description
The data described in this article consist of all Nature-Inspired Algorithms that have been published to-date. To define which meta-heuristics can be considered Nature-Inspired, the definition given by [1] is used, stating that "the term nature refers to any part of the physical universe which is not a product of intentional human design".
The database consists of 43 variables, as follows: The data include the algorithm's name (variable 1), the abbreviation (variable 2), the year presented (variable 3), the authors (variables 6-14, where applicable), the Journal or Conference where the algorithm was published initially (variables 15-16), and the applications tackled in this initial work (variables 19-22) ( Table 1 ).
In variables 4 and 5, the algorithms are categorised based on their main inspiration category and the sub-category, as in [ 2 , 3 ]. The categories and the subcategories are presented in Table 2: Variable 15 categorises algorithms based on where they were presented as: Text First author of initial work presenting the algorithm 7 Author2 Text Second author of initial work presenting the algorithm 8 Author3 Text Third author of initial work presenting the algorithm 9 Author4 Text Next author of initial work presenting the algorithm 10 Author5 Text Next author of initial work presenting the algorithm 11 Author6 Text Next author of initial work presenting the algorithm 12 Author7 Text Next author of initial work presenting the algorithm 13 Author8 Text Next author of initial work presenting the algorithm 14 Author9 Text Next author of initial work presenting the algorithm 15 Publication Categorical Where was the algorithm presented initially (Journal or Conference)?    Table 3 .
All applications areas, where at least one algorithm has been applied, are given in Table 4 .
Variable 18 provides an algorithm taxonomy based on the application tackled in the initial work, as:

No 2 Yes 3 Only Benchmark functions
Notes about the algorithm are included in variable 23. In the initial version of the dataset only one algorithm has a note, which initial work has been retracted. This note has taken the value of 1, and in future versions of the data set, more values would be added if applicable.
Furthermore, the data include the number of published papers in five optimisation problem areas, i.e. engineering problems (variables 24-25), financial problems (variables 26-27), op- Linköping University Electronic Press erations research (variables 28-29), energy problems (variables 30-31) and other optimisation problems (variables 32-33). The total number of all these areas can be seen in the corresponding feature (variables 34-35), As well asapplications of each algorithm in clustering and/or classification problems (variables 36-37), and also forecasting ones (variables 38-39). In all cases, two variables are used, where the first variable of each pair denotes the total number of works, while the second one denotes only the number of works published in Journals.

Experimental design, materials, and methods
The data described in this article have been acquired from 2017 todate. They are divided into preliminary data acquired through documentation (variables 1-23) and secondary data (variables 24-44), which have been calculated using several scientific repositories.
Initially, based on the work of [4] , the authors collected some Nature-Inspired algorithms, where the algorithm's name, the abbreviation, the year presented, the authors, the Journal or Conference where the algorithm was initially published and the applications tackled in this initial work have been noted. IBM's SPSS package was used to organise all these features. Useful information has also been found in [ 5 , 6 ]. This database has been updated on a monthly basis.
Furthermore, the number of papers where each algorithm is applied in various problem areas has been calculated through web research. Using Google Scholar, Mendeley and other scientific repositories, [ 2 , 3 ], we collected the number of published papers in five optimisation problem areas, i.e. engineering problems, financial problems, operations research, energy problems and other optimisation problems. The total number of applications of each algorithm has been calculated in optimisation, clustering and/or classification problems, while also forecasting problems.
The total works have been calculated from the above information, and a binary variable denotes if the algorithms have been applied in at least one problem area, without taking into consideration the work where they were published. Another categorical variable has been added, which performs taxonomy of the algorithms based on the total published works. Based on the number of applications, a dummy variable has been generated, in which algorithms are classified as non-established and established algorithms.
The monthly update is performed via web research in scientific repositories, as is shown in Fig. 1 . The final database is exported in csv format.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships which have, or could be perceived to have, influenced the work reported in this article.