Bradley Alexander

Evolving patches for software repair

Defects are a major concern in software systems. Unsurprisingly, there are many tools and techniques to facilitate the removal of defects through their detection and localisation. However, there are few tools that attempt to repair defects. To date, evolutionary tools for software repair have evolved changes directly in the program code being repaired. In this work we describe an implementation: pyEDB, that encodes changes as a series of code modifications or patches. These modifications are evolved as individuals. We show pyEDB to be effective in repairing some small errors, including variable naming errors in Python programs. We also demonstrate that evolving patches rather than whole programs simplifies the removal of spurious errors.

Reformulated Co-Tree Flows Method Competitive with the Global Gradient Algorithm for Solving Water Distribution System Equations

AbstractMany different methods have been devised to solve the nonlinear systems of equations that model water distribution networks. Probably the most popular is Todini and Pilati’s global gradient algorithm (GGA). Given the GGA’s success, alternative methods have not aroused much interest. One example is the co-tree method, which requires some cumbersome steps in its implementation. In this paper, a reformulated co-trees method (RCTM) is presented that simplifies the procedure by manipulating the incidence matrix into trapezoidal form: a lower triangular block at the top representing a spanning tree and rectangular block below it representing the corresponding co-tree. This reordering leads to significant efficiencies that make the RCTM competitive with the GGA in certain settings. The new method has some similarities to the loop flows corrections formulation, and it is shown, by application to a set of eight case study networks with between 932 and 19,647 pipes and between 848 and 17,971 nodes, to be be...

Evolutionary Image Transition Using Random Walks

We present a study demonstrating how random walk algorithms can be used for evolutionary image transition. We design different mutation operators based on uniform and biased random walks and study how their combination with a baseline mutation operator can lead to interesting image transition processes in terms of visual effects and artistic features. Using feature-based analysis we investigate the evolutionary image transition behaviour with respect to different features and evaluate the images constructed during the image transition process.

The Evolutionary Process of Image Transition in Conjunction with Box and Strip Mutation

Evolutionary algorithms have been used in many ways to generate digital art. We study how evolutionary processes are used for evolutionary art and present a new approach to the transition of images. Our main idea is to define evolutionary processes for digital image transition, combining different variants of mutation and evolutionary mechanisms. We introduce box and strip mutation operators which are specifically designed for image transition. Our experimental results show that the process of an evolutionary algorithm in combination with these mutation operators can be used as a valuable way to produce unique generative art.

Evolution of artistic image variants through feature based diversity optimisation

Measures aimed to improve the diversity of images and image features in evolutionary art help to direct search toward more novel and creative parts of the artistic search domain. To date such measures have not focused on selecting from all individuals based on their contribution to diversity of feature metrics. In recent work on TSP problem instance classification, selection based on a direct measure of each individuals contribution to diversity was successfully used to generate hard and easy TSP instances. In this work we use this search framework to evolve diverse variants of a source image in one and two feature dimensions. The resulting images show the spectrum of effects from transforming images to score across the range of each feature. The results also reveal interesting correlations between feature values in two dimensions.

Evolving a path planner for a multi-robot exploration system using grammatical evolution

Area exploration and mapping with teams of robots is a challenging application. As the complexity of this application increases so does the challenge of designing effective coordinated control. One potential solution to this problem is to explore some relevant parts of the design space automatically. In this paper, we present an approach which uses Grammatical Evolution to design a control function for coordinated path planning of teams of mobile robots. Simulation results are promising with evolved control functions showing performance better than handwritten control in term of amount of explored area.

Deep parameter optimisation on Android smartphones for energy minimisation: a tale of woe and a proof-of-concept

With power demands of mobile devices rising, it is becoming increasingly important to make mobile software applications more energy efficient. Unfortunately, mobile platforms are diverse and very complex which makes energy behaviours difficult to model. This complexity presents challenges to the effectiveness of off-line optimisation of mobile applications. In this paper, we demonstrate that it is possible to automatically optimise an application for energy on a mobile device by evaluating energy consumption in-vivo. In contrast to previous work, we use only the devices own internal meter. Our approach involves many technical challenges but represents a realistic path toward learning hardware specific energy models for program code features.

Evolving decision-making functions in an autonomous robotic exploration strategy using grammatical evolution

Customising navigational control for autonomous robotic mapping platforms is still a challenging task. Control software must simultaneously maximise the area explored whilst maintaining safety and working within the constraints of the platform. Scoring functions to assess navigational options are typically written by hand and manually refined. As navigational tasks become more complex this manual approach is unlikely to yield the best results. In this paper we explore the automatic derivation of a scoring function for a ground based exploration platform. We show that it is possible to derive the entire structure of a scoring function and that allowing structure to evolve yields significant performance advantages over the evolution of embedded constants alone.

A detailed comparison of meta-heuristic methods for optimising wave energy converter placements

In order to address environmental concerns and meet growing energy demand the development of green energy technology has expanded tremendously. One of the most promising types of renewable energy is ocean wave energy. While there has been strong research in the development of this technology to date there remain a number of technical hurdles to overcome. This research explores a type of wave energy converter (WEC) called a buoy. This work models a power station as an array of fully submerged three-tether buoys. The target problem of this work is to place buoys in a size-constrained environment to maximise power output. This article improves prior work by using a more detailed model and exploring the search space using a wide variety of search heuristics. We show that a hybrid method of stochastic local search combined with Nelder-Mead Simplex direct search performs better than previous search techniques.

Application of evolutionary methods to 3D geoscience modelling

Geoscience modelling plays a vital role in mapping and tracking the earths resources. Magnetotellurics, which maps the electrical resistivity of the subsurface, is a useful and cost-effective sounding-technique for sensing over a broad scale at depth. However, due to the inherent difficulty in sensing at depth, models produced using MT have a degree of uncertainty. Geoscientists can reduce this uncertainty by producing multiple alternative models, and using multiple modelling techniques and settings, to correlate robust model features with field data responses. Population-based evolutionary search techniques are of interest to MT modelling because they offer an alternative to deterministic techniques, and are able to produce multiple models for analysis. Unfortunately, evolutionary techniques have not been successfully applied to 3D MT modelling. In this work we describe a new, more compact, representation of MT models using volumetric functions. Using this representation we successfully apply evolutionary search techniques to 3D MT modelling for both artificial and real models and show how the development of large scale features during modelling can be correlated with the models fit to field data.