Evolutionary algorithms (EAs) are general-purpose optimizers that mimick principles from the natural evolution of species. They maintain a collection of possible solutions (the population) and then apply operators like mutation and/or recombination to create new solutions (the offspring). A selection process then chooses a new population for the next generation. Evolutionary algorithms are popular in practice as they can be easily applied to various problems. In contrast to problem-specific

more »

... rithms they require minimal or no knowledge about the problem in hand, as long as there is a way to quantify the solution quality (fitness) of new solutions. A challenge when applying EAs is that their performance and their dynamic behavior is not well understood. To remedy this, theoretical computer scientists employ methods from the analysis of randomized algorithms to analyze the performance of EAs with mathematical rigor. The aim is to develop a fundamental understanding of these algorithms and to aid in the design of new and more effective EAs. The theory track of the annual ACM Genetic and Evolutionary Computation Conference (GECCO) is the first tier event for advances in this direction. In this special issue nine selected papers from the 2017 edition of the GECCO theory track are collected, each one of them carefully revised and extended to meet the high quality standards of Algorithmica. Many evolutionary algorithms use an offspring population; i.e., in each iteration they evaluate the quality of several solution candidates, so as to profit from paralleling their quality assessment. The quality of these search points determines where and how the search is continued. One of the simplest algorithms using such a population-based approach is the (1+λ) Evolutionary Algorithm, shortly (1+λ) EA. It is known that EAs with large offspring populations can benefit from an increased mutation rate as λ offspring can amplify the chances of making good progress in one generation. But B Carola Doerr