Alex Shenfield

Evolutionary determination of experimental parameters for ptychographical imaging

The Ptychographical Iterative Engine (PIE) algorithm is a recently developed novel method of Coherent Diffractive Imaging (CDI) that uses multiple overlapping diffraction patterns to reconstruct an image. This method has successfully produced high quality reconstructions at both optical and X-ray wavelengths but the need for accurate knowledge of the probe positions is currently a limiting factor in the production of high resolution reconstructions at electron wavelengths. This paper examines the shape of the search landscape for producing optimal image reconstructions in the specific case of electron microscopy and then shows how evolutionary search methods can be used to reliably determine experimental parameters in the electron microscopy case (such as the spherical aberration in the probe and the probe positions).

A novel preference articulation operator for the Evolutionary Multi-Objective Optimisation of classifiers in concealed weapons detection

The incorporation of decision maker preferences is often neglected in the Evolutionary Multi-Objective Optimisation (EMO) literature. The majority of the research in the field and the development of EMO algorithms is primarily focussed on converging to a Pareto optimal approximation close to or along the true Pareto front of synthetic test problems. However, when EMO is applied to real-world optimisation problems there is often a decision maker who is only interested in a portion of the Pareto front (the Region of Interest) which is defined by their expressed preferences for the problem objectives. In this paper a novel preference articulation operator for EMO algorithms is introduced (named the Weighted Z-score Preference Articulation Operator) with the flexibility of being incorporated a priori, a posteriori or progressively, and as either a primary or auxiliary fitness operator. The Weighted Z-score Preference Articulation Operator is incorporated into an implementation of the Multi-Objective Evolutionary Algorithm Based on Decomposition (named WZ-MOEA/D) and benchmarked against MOEA/D-DRA on a number of bi-objective and five-objective test problems with test cases containing preference information. After promising results are obtained when comparing WZ-MOEA/D to MOEA/D-DRA in the presence of decision maker preferences, WZ-MOEA/D is successfully applied to a real-world optimisation problem to optimise a classifier for concealed weapon detection, producing better results than previously published classifier implementations.

Computational steering of a multi-objective evolutionary algorithm for engineering design

The execution process of an evolutionary algorithm typically involves some trial and error. This is due to the difficulty in setting the initial parameters of the algorithm-especially when little is known about the problem domain. This problem is magnified when applied to many-objective optimisation, as care is needed to ensure that the final population of candidate solutions is representative of the trade-off surface. We propose a computational steering system that allows the engineer to interact with the optimisation routine during execution. This interaction can be as simple as monitoring the values of some parameters during the execution process, or could involve altering those parameters to influence the quality of the solutions produced by the optimisation process. The implementation of this steering system should provide the ability to tailor the client to the hardware available, for example providing a lightweight steering and visualisation client for use on a PDA.

Optimisation of maintenance scheduling strategies on the grid

The emerging paradigm of Grid Computing provides a powerful platform for the optimisation of complex computer models, such as those used to simulate real-world logistics and supply chain operations. This paper introduces a Grid-based optimisation framework that provides a powerful tool for the optimisation of such computationally intensive objective functions. This framework is then used in the optimisation of maintenance scheduling strategies for fleets of aero-engines, a computationally intensive problem with a high-degree of stochastic noise, achieving substantial improvements in the execution time of the algorithm.

Multi-objective evolutionary design of robust controllers on the grid

Coupling conventional controller design methods, model based controller synthesis and simulation, and multi-objective evolutionary optimisation methods frequently results in an extremely computationally expensive design process. However, the emerging paradigm of grid computing provides a powerful platform for the solution of such problems by providing transparent access to large-scale distributed high-performance compute resources. As well as substantially speeding up the time taken to find a single controller design satisfying a set of performance requirements this grid-enabled design process allows a designer to effectively explore the solution space of potential candidate solutions. An example of this is in the multi-objective evolutionary design of robust controllers, where each candidate controller design has to be synthesised and the resulting performance of the compensated system evaluated by computer simulation. This paper introduces a grid-enabled framework for the multi-objective optimisation of computationally expensive problems which will then be demonstrated using and example of the multi-objective evolutionary design of a robust lateral stability controller for a real-world aircraft using H ∞ loop shaping.

CMA-PAES: Pareto archived evolution strategy using covariance matrix adaptation for Multi-Objective Optimisation

The quality of Evolutionary Multi-Objective Optimisation (EMO) approximation sets can be measured by their proximity, diversity and pertinence. In this paper we introduce a modular and extensible Multi-Objective Evolutionary Algorithm (MOEA) capable of converging to the Pareto-optimal front in a minimal number of function evaluations and producing a diverse approximation set. This algorithm, called the Covariance Matrix Adaptation Pareto Archived Evolution Strategy (CMA-PAES), is a form of (μ + λ) Evolution Strategy which uses an online archive of previously found Pareto-optimal solutions (maintained by a bounded Pareto-archiving scheme) as well as a population of solutions which are subjected to variation using Covariance Matrix Adaptation. The performance of CMA-PAES is compared to NSGA-II (currently considered the benchmark MOEA in the literature) on the ZDT test suite of bi-objective optimisation problems and the significance of the results are analysed using randomisation testing.

A service oriented architecture for decision making in engineering design

Decision making in engineering design can be effectively addressed by using genetic algorithms to solve multi-objective problems. These multi-objective genetic algorithms (MOGAs) are well suited to implementation in a Service Oriented Architecture. Often the evaluation process of the MOGA is compute-intensive due to the use of a complex computer model to represent the real-world system. The emerging paradigm of Grid Computing offers a potential solution to the compute-intensive nature of this objective function evaluation, by allowing access to large amounts of compute resources in a distributed manner. This paper presents a grid-enabled framework for multi-objective optimisation using genetic algorithms (MOGA-G) to aid decision making in engineering design.

A Novel Workload Allocation Strategy for Batch Jobs

The distribution of computational tasks across a diverse set of geographically distributed heterogeneous resources is a critical issue in the realisation of true computational grids. Conventionally, workload allocation algorithms are divided into static and dynamic approaches. Whilst dynamic approaches frequently outperform static schemes, they usually require the collection and processing of detailed system information at frequent intervals - a task that can be both time consuming and unreliable in the real-world. This paper introduces a novel workload allocation algorithm for optimally distributing the workload produced by the arrival of batches of jobs. Results show that, for the arrival of batches of jobs, this workload allocation algorithm outperforms other commonly used algorithms in the static case. A hybrid scheduling approach (using this workload allocation algorithm), where information about the speed of computational resources is inferred from previously completed jobs, is then introduced and the efficiency of this approach demonstrated using a real world computational grid. These results are compared to the same workload allocation algorithm used in the static case and it can be seen that this hybrid approach comprehensively outperforms the static approach.

Automated detection of atrial fibrillation using long short-term memory network with RR interval signals

Atrial Fibrillation (AF), either permanent or intermittent (paroxysnal AF), increases the risk of cardioembolic stroke. Accurate diagnosis of AF is obligatory for initiation of effective treatment to prevent stroke. Long term cardiac monitoring improves the likelihood of diagnosing paroxysmal AF. We used a deep learning system to detect AF beats in Heart Rate (HR) signals. The data was partitioned with a sliding window of 100 beats. The resulting signal blocks were directly fed into a deep Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM). The system was validated and tested with data from the MIT-BIH Atrial Fibrillation Database. It achieved 98.51% accuracy with 10-fold cross-validation (20 subjects) and 99.77% with blindfold validation (3 subjects). The proposed system structure is straight forward, because there is no need for information reduction through feature extraction. All the complexity resides in the deep learning system, which gets the entire information from a signal block. This setup leads to the robust performance for unknown data, as measured with the blind fold validation. The proposed Computer-Aided Diagnosis (CAD) system can be used for long-term monitoring of the human heart. To the best of our knowledge, the proposed system is the first to incorporate deep learning for AF beat detection.

A multi objective approach to evolving artificial neural networks for coronary heart disease classification

The optimisation of the accuracy of classifiers in pattern recognition is a complex problem that is often poorly understood. Whilst numerous techniques exist for the optimisation of weights in artificial neural networks (e.g. the Widrow-Hoff least mean squares algorithm and back propagation techniques), there do not exist any hard and fast rules for choosing the structure of an artificial neural network - in particular for choosing both the number of the hidden layers used in the network and the size (in terms of number of neurons) of those hidden layers. However, this internal structure is one of the key factors in determining the accuracy of the classification. This paper proposes taking a multi-objective approach to the evolutionary design of artificial neural networks using a powerful optimiser based around the state-of-the-art MOEA/D-DRA algorithm and a novel method of incorporating decision maker preferences. In contrast to previous approaches, the novel approach outlined in this paper allows the intuitive consideration of trade-offs between classification objectives that are frequently present in complex classification problems but are often ignored. The effectiveness of the proposed multi-objective approach to evolving artificial neural networks is then shown on a real-world medical classification problem frequently used to benchmark classification methods.