Lukas König

Decentralized evolution of robotic behavior using finite state machines

Purpose – In evolutionary robotics (ER), robotic control systems are subject to a developmental process inspired by natural evolution. The purpose of this paper is to utilize a control system representation based on finite state machines (FSMs) to build a decentralized online‐evolutionary framework for swarms of mobile robots.Design/methodology/approach – A new recombination operator for multi‐parental generation of offspring is presented and a known mutation operator is extended to harden parts of genotypes involved in good behavior, thus narrowing down the dimensions of the search space. A storage called memory genome for archiving the best genomes of every robot introduces a decentralized elitist strategy. These operators are studied in a factorial set of experiments by evolving two different benchmark behaviors such as collision avoidance and gate passing on a simulated swarm of robots. A comparison with a related approach is provided.Findings – The framework is capable of robustly evolving the benchm...

A Completely Evolvable Genotype-Phenotype Mapping for Evolutionary Robotics

To achieve a desired global behavior for a swarm of robots where each robot has a local view and operating range in the environment is a well-known and challenging problem. Evolutionary Robotics is a self-adaptation approach which has been shown to e effectively find robot controllers for behaviors which are hard to implement by hand. There, evolvability is highly dependent on controller representation during evolution. It is known that using a genotypic controller representation which also encodes parts of the genotype-phenotype mapping (GPM) can lead to a meta-adaptation of the evolutionary operators to the search space structure, thus improving evolvability. We enhance this idea using a fully flexible GPM which is represented in the same way as the behavioral controllers are, and, therefore, can be completely evolved along with the behavior. The approach is based on finite state machines and extends an existing framework for decentralized evolution of robot behavior in swarms of mobile robots. Experiments indicate that the evolvable GPM outperforms both the extensively improved operators of the existing framework and a standard operator for the new real-valued genotypes with fixed GPM.

Evolving Collision Avoidance on Autonomous Robots

Utilizing the collective behavior of a population of interacting individuals, based on rather simple local algorithms, is a promising approach for achieving complex goals. We use an onboard online evolutionary model, based on finite Moore automata, to develop collective behavior in an artificial swarm of micro-robots. Experiments have been made in simulation to achieve Collision Avoidance. The model is shown to be capable to generate the desired behavior and we present experiments for adjusting the parameters of the evolutionary optimization.

Stigmergy-Based Scheduling of Flexible Loads

In this paper, we address the rescheduling of shiftable loads in a sub-section of the power grid (micro-grid) to maximize the utilization of renewable energy sources (RES). The objective is to achieve a schedule for all customers in the micro-grid such that the RES output utilization is maximized. Customers correspond to residential households provided with intelligent appliances with the ability to recalculate their operation times. We propose an approach based on stigmergy to efficiently find a close-to-optimal solution to the general problem. An empirical analysis of the internal functioning of the algorithm is performed. Furthermore, the performance of the algorithm is compared to a price-based approach.

Introducing the simulation plugin interface and the EAS framework with comparison to two state-of-the-art agent simulation frameworks

This paper proposes a novel architectural concept for developing agent-based simulations called Simulation Plugin Interface (SPI); furthermore, a simulation framework called Easy Agent Simulation (EAS) based on the proposed architecture is presented. The SPI introduces an intermediate layer between the simulation engine and the simulation model. It contains all types of functionality which are required for a simulation but are logically separable from the simulation model. This includes visualization, probes, statistics calculations, logging, scheduling, API to other programming languages, etc. The architecture is particularly suitable to guide student programmers with low experience to well-structured and reusable simulation components. The SPI architecture is not bound to the EAS Framework, but can be implemented as an extension to most state-of-the-art simulation frameworks. In a comparative study, the EAS framework is compared to the agent simulation frameworks NetLogo and MASON, using the well-known “Stupid Model” as a test scenario.