Self-Replicating Robot System Simulation Data using basic iterative production algorithm

Published: 6 January 2021| Version 1 | DOI: 10.17632/hnvh4zmg5p.1
Andrew Jones


The dataset contains the results of 100 simulation runs for each of six different production approach algorithms for self-replicating robot systems under 70 different experimental conditions. For each production approach under each experimental condition, the data contains the assemble, print, and collection capability values along with their standard deviation values. It also includes the values for the average robot build quality, number of robots destroyed, and the number of robots being assembled at the end of the simulation run. This data can be used to compare the efficacy of the six different production approach algorithms.


Steps to reproduce

Using the self-replicating robot foraging and production decision-making simulation system, change the parameters from the defaults to those of the condition that is listed in the ID_KEY sheet of the excel file (included in the data download). Then, run the simulator and results will be output as a CSV file.


North Dakota State University


Artificial Intelligence, Robotics, Computer Simulation, Three Dimensional Printing