Evan J. Hughes

Multiple single objective Pareto sampling

We detail a new nonPareto evolutionary multiobjective algorithm, multiple single objective Pareto sampling (MSOPS), that performs a parallel search of multiple conventional target vector based optimisations, e.g. weighted min-max. The method can be used to generate the Pareto set and analyse problems with large numbers of objectives. The method allows bounds and discontinuities of the Pareto set to be identified and the shape of the surface to be analysed, despite not being able to visualise the surface easily. A new combination metric is also introduced that allows the shape of the objective surface that gives rise to discontinuities in the Pareto surface to be analysed easily.

MSOPS-II: A general-purpose Many-Objective optimiser

Existing evolutionary methods capable of true many-objective optimisation have been limited in their application: for example either initial search directions need to be specified a-priori, or the use of hypervolume limits the search in practice to less than 10 objective dimensions. This paper describes two extensions to the multiple single objective pareto sampling (MSOPS) algorithm. The first provides automatic target vector generation, removing the requirement for initial a-priori designer intervention; and secondly redefines the fitness assignment method to simplify analysis and allow more comprehensive constraint handling. The significant enhancements allow the new MSOPS-II ranking process to be used as part of a general-purpose multi/many objective optimisation algorithm, requiring minimal initial configuration.

Radar waveform optimisation as a many-objective application benchmark

This paper introduces a real, unmodified Many-Objective optimisation problem for use in optimisation algorithm benchmarking. The radar waveform design problem has 9 objectives and an integer decision space that can be scaled from 4 to 12 decision variables. Proprietary radar waveform design software has been encapsulated in a fast and portable form to facilitate research groups in studying high-order optimisation of real engineering problems.

Medium PRF radar PRF selection using evolutionary algorithms

Evolutionary algorithms are applied to the optimization of pulse repetition frequency (PRF), for both eight-and nine PRFs, in medium PRF radar while considering the detailed effects of sidelobe clutter and many other technical factors. The algorithm presented also ensures that all the solutions produced are fully decodable and have no blind velocities. The evolutionary algorithm was able to identify near-optimum PRF sets for a realistic radar system with only a modest computational effort.

Fitness Assignment Methods for Many-Objective Problems

This chapter considers a number of alternative methods for fitness assignment in evolutionary algorithms for multiobjective optimization. Most of the fitness assignment methods in the literature were designed to work for any number of objectives, in principle; but, in practice, some of the more popular methods (e.g. those in NSGA-II, IBEA and SPEA) do not perform well on problems with four or more objectives. We investigate why this is the case, considering two aspects of performance: convergence towards the Pareto front and drive towards a set of well spread solutions. The visualization of induced fitness surfaces is used to understand the effects of the different fitness assignment methods, and both Pareto- and non-Pareto-based methods are analysed.

Medium PRF set selection using evolutionary algorithms

This paper presents a new and novel method of selecting multiple pulse repetition frequency (PRF) sets for use in medium PRF pulsed-Doppler radars. Evolutionary algorithms are used to minimise the blind areas in the range/Doppler space. The evolutionary algorithm allows optimal solutions to be generated quickly, far faster than with exhaustive searches, and is fully automatic, unlike existing techniques. The evolved solutions compare very favorably against the results of both an exhaustive search and existing published PRF set selection methods. This evolutionary approach to generation of PRF sets is a major advance in medium PRF radar design.

Fuzzy autopilot design using a multiobjective evolutionary algorithm

This paper details a Fuzzy-Feedback Linearisation controller applied to a non-linear missile. The design uses an evolutionary algorithm optimisation approach to a multiple model description of the airframe aerodynamics. A set of convex models is produced that map the vertex points in a high order parameter space (of the order of 16 variables). These are used to determine the membership function distribution within the outer loop control system by using a multi-objective evolutionary algorithm. This produces a design that meets objectives related to closed loop performance such as: steady state error, overshoot, settling and rising time. The evolutionary algorithm uses non-dominated sorting for forming a Pareto front of possible solutions. This paper shows that fuzzy controllers can be produced for engineering problems, with the multiobjective algorithm allowing the designer the freedom to choose solutions and investigate the properties of the system.

Using multiple genetic algorithms to generate radar point-scatterer models

This paper covers the use of three different genetic algorithms applied sequentially to radar cross-section data to generate point-scatterer models. The aim is to provide automatic conversion of measured 2D/3D data of low, medium, or, high resolution into scatterer models. The resulting models are intended for use in a missile-target engagement simulator. The first genetic algorithm uses multiple species to locate the scattering centers. The second and third algorithms are for model fine tuning and optimization, respectively. Both of these algorithms use nondominated ranking to generate Pareto-optimal sets of results. The ability to choose results from the Pareto sets allows the designer some flexibility in the creation of the model. A method for constructing compound models to produce full 4 /spl pi/ sr coverage is detailed. Example results from the model generation process are presented.

Many-objective directed evolutionary line search

Algorithms capable of performing efficient and controllable many objective optimisation are becoming more necessary as the complexity of optimisation problems to be solved increases. This paper describes a new algorithm that combines elements of traditional gradient based optimisation methods along with a powerful many-objective capable search process. The algorithm exploits the directed line search (such as Golden Section Search) procedures found in many single-objective gradient based algorithms in order to both explore and exploit features in the optimisation landscape. The target vector and aggregation methods used in the MSOPS algorithm have been employed to provide effective and controllable many-objective optimisation, especially suited to close interaction with a designer where it is often desired to target specific regions of the Pareto front. The Many Objective Directed Evolutionary Line Search (MODELS) algorithm is demonstrated on a constrained function with a concave Pareto front in up to 20 dimensions and is shown to outperform existing optimisers, some of which are known to perform well for many-objective problems.

Performance comparison of PRF schedules for medium PRF radar

Previous work has shown how evolutionary algorithms (EAs) are an effective tool in optimising the selection of pulse repetition frequency (PRF) values of medium PRF schedules in an airborne fire control radar (FCR) application requiring target data in three PRFs. The optimisation is driven by the requirement to minimise range/Doppler blindness whilst maintaining full decodability. In this paper we detail work in which the optimisation process is applied to design novel short medium PRF schedules requiring target data in just two PRFs. The paper reports on the testing of a variety of near-optimum schedules to compare their blindness, decoding, and ghosting performances. The results show that in many situations, the 2 of N schedules are a practical alternative to conventional 3 of N processing.