Evolutionary Informatics

Research Tools

The Evolutionary Informatics Lab presents the free online tools below to allow you to investigate aspects of evolutionary algorithms from the comfort of your own browser.

• Minivida

  

  Description

  Minivida is a simplified, online simulation of Lenski et. al.'s Avida. It allows the evolution and visualization of NAND-logic programs. With it, you can learn how the choice of rewarded tasks and instructions affects the ability of locating EQU operations. Does the process of mutation and selection alone allow the evolution of complex operations or is the presence of stair step information necessary in order to guide the process? Run the simulation and see what role user-supplied information plays in the discovery of complex features.

• Ev Ware

  

  Description

  Ev Ware is a simulation of T.M. Schneider's ev evolutionary binding site simulator. Our simulation aims to be a faithful replication of ev and produces behavior representative of the original program. Ev Ware features a multi-run lab, allowing you to perform multiple automated experiments on the behavior of the ev perceptron. With Ev Ware, you can test and limit how much information is imposed on the search by the structure of the program and by particular parameter settings. Discover how likely randomized initializations are to stumble upon targets of your choice. Does ev create information from “scratch” or does it merely reshuffle pre-supplied information? Find out. Usage instructions are included and source code is available.

• Weasel Ware

  

  Description

  A lot of discussion has surrounded Dawkins’ Weasel evolutionary simulation since he introduced it in his book The Blind Watchmaker. Weasel Ware is a browser-based simulation of Dawkins’ Weasel, covering the relevant aspects of his model and allowing analysis of simple evolutionary algorithms containing target-specific, information-rich oracles. Do the mere presence of replication, mutation and selection guarantee success in a search? If not, what fraction of all possible fitness functions lead to success? How likely are we to stumble across a successful fitness function if we initialize one at random? The Weasel Ware research tool allows users to answer these questions and more. Simply set up your experiment, choose your fitness function and record your results. Additionally, users can design and use fitness functions of their own, through the custom fitness function mode..