Using ancillary statistics in on-line learning algorithms

Neural networks are usually curved statistical models. They do not have finite dimensional sufficient statistics, so on-line learning on the model itself inevitably loses information. In this paper we propose a new scheme for training curved models, inspired by the ideas of ancillary statistics and adaptive critics. At each point estimate an auxiliary flat model (exponential family) is built to locally accommodate both the usual statistic (tangent to the model) and an ancillary statistic (normal to the model). The auxiliary model plays a role in determining credit assignment analogous to that played by an adaptive critic in solving temporal problems. The method is illustrated with the Cauchy model and the algorithm is proved to be asymptotically efficient.

Modelling the dynamics of genetic algorithms using statistical mechanics

A formalism for modelling the dynamics of Genetic Algorithms (GAs) using methods from statistical mechanics, originally due to Prugel-Bennett and Shapiro, is reviewed, generalized and improved upon. This formalism can be used to predict the averaged trajectory of macroscopic statistics describing the GA's population. These macroscopics are chosen to average well between runs, so that fluctuations from mean behaviour can often be neglected. Where necessary, non-trivial terms are determined by assuming maximum entropy with constraints on known macroscopics. Problems of realistic size are described in compact form and finite population effects are included, often proving to be of fundamental importance. The macroscopics used here are cumulants of an appropriate quantity within the population and the mean correlation (Hamming distance) within the population. Including the correlation as an explicit macroscopic provides a significant improvement over the original formulation. The formalism is applied to a number of simple optimization problems in order to determine its predictive power and to gain insight into GA dynamics. Problems which are most amenable to analysis come from the class where alleles within the genotype contribute additively to the phenotype. This class can be treated with some generality, including problems with inhomogeneous contributions from each site, non-linear or noisy fitness measures, simple diploid representations and temporally varying fitness. The results can also be applied to a simple learning problem, generalization in a binary perceptron, and a limit is identified for which the optimal training batch size can be determined for this problem. The theory is compared to averaged results from a real GA in each case, showing excellent agreement if the maximum entropy principle holds. Some situations where this approximation brakes down are identified. In order to fully test the formalism, an attempt is made on the strong sc np-hard problem of storing random patterns in a binary perceptron. Here, the relationship between the genotype and phenotype (training error) is strongly non-linear. Mutation is modelled under the assumption that perceptron configurations are typical of perceptrons with a given training error. Unfortunately, this assumption does not provide a good approximation in general. It is conjectured that perceptron configurations would have to be constrained by other statistics in order to accurately model mutation for this problem. Issues arising from this study are discussed in conclusion and some possible areas of further research are outlined.

Magnification factors for the SOM and GTM algorithms

Magnification factors specify the extent to which the area of a small patch of the latent (or `feature') space of a topographic mapping is magnified on projection to the data space, and are of considerable interest in both neuro-biological and data analysis contexts. Previous attempts to consider magnification factors for the self-organizing map (SOM) algorithm have been hindered because the mapping is only defined at discrete points (given by the reference vectors). In this paper we consider the batch version of SOM, for which a continuous mapping can be defined, as well as the Generative Topographic Mapping (GTM) algorithm of Bishop et al. (1997) which has been introduced as a probabilistic formulation of the SOM. We show how the techniques of differential geometry can be used to determine magnification factors as continuous functions of the latent space coordinates. The results are illustrated here using a problem involving the identification of crab species from morphological data.

Noisy fitness evaluation in genetic algorithms and the dynamics of learning

A theoretical model is presented which describes selection in a genetic algorithm (GA) under a stochastic fitness measure and correctly accounts for finite population effects. Although this model describes a number of selection schemes, we only consider Boltzmann selection in detail here as results for this form of selection are particularly transparent when fitness is corrupted by additive Gaussian noise. Finite population effects are shown to be of fundamental importance in this case, as the noise has no effect in the infinite population limit. In the limit of weak selection we show how the effects of any Gaussian noise can be removed by increasing the population size appropriately. The theory is tested on two closely related problems: the one-max problem corrupted by Gaussian noise and generalization in a perceptron with binary weights. The averaged dynamics can be accurately modelled for both problems using a formalism which describes the dynamics of the GA using methods from statistical mechanics. The second problem is a simple example of a learning problem and by considering this problem we show how the accurate characterization of noise in the fitness evaluation may be relevant in machine learning. The training error (negative fitness) is the number of misclassified training examples in a batch and can be considered as a noisy version of the generalization error if an independent batch is used for each evaluation. The noise is due to the finite batch size and in the limit of large problem size and weak selection we show how the effect of this noise can be removed by increasing the population size. This allows the optimal batch size to be determined, which minimizes computation time as well as the total number of training examples required.

Optimization of physical distribution problem in logistics management

Physical distribution plays an imporant role in contemporary logistics management. Both satisfaction level of of customer and competitiveness of company can be enhanced if the distribution problem is solved optimally. The multi-depot vehicle routing problem (MDVRP) belongs to a practical logistics distribution problem, which consists of three critical issues: customer assignment, customer routing, and vehicle sequencing. According to the literatures, the solution approaches for the MDVRP are not satisfactory because some unrealistic assumptions were made on the first sub-problem of the MDVRP, ot the customer assignment problem. To refine the approaches, the focus of this paper is confined to this problem only. This paper formulates the customer assignment problem as a minimax-type integer linear programming model with the objective of minimizing the cycle time of the depots where setup times are explicitly considered. Since the model is proven to be MP-complete, a genetic algorithm is developed for solving the problem. The efficiency and effectiveness of the genetic algorithm are illustrated by a numerical example.

An integrated scheduling problem of PCB components on sequential pick-and-place machines:mathematical models and heuristic solutions

This paper formulates several mathematical models for determining the optimal sequence of component placements and assignment of component types to feeders simultaneously or the integrated scheduling problem for a type of surface mount technology placement machines, called the sequential pick-andplace (PAP) machine. A PAP machine has multiple stationary feeders storing components, a stationary working table holding a printed circuit board (PCB), and a movable placement head to pick up components from feeders and place them to a board. The objective of integrated problem is to minimize the total distance traveled by the placement head. Two integer nonlinear programming models are formulated first. Then, each of them is equivalently converted into an integer linear type. The models for the integrated problem are verified by two commercial packages. In addition, a hybrid genetic algorithm previously developed by the authors is adopted to solve the models. The algorithm not only generates the optimal solutions quickly for small-sized problems, but also outperforms the genetic algorithms developed by other researchers in terms of total traveling distance.

A genetic algorithm for the generalised transportation problem

The generalised transportation problem (GTP) is an extension of the linear Hitchcock transportation problem. However, it does not have the unimodularity property, which means the linear programming solution (like the simplex method) cannot guarantee to be integer. This is a major difference between the GTP and the Hitchcock transportation problem. Although some special algorithms, such as the generalised stepping-stone method, have been developed, but they are based on the linear programming model and the integer solution requirement of the GTP is relaxed. This paper proposes a genetic algorithm (GA) to solve the GTP and a numerical example is presented to show the algorithm and its efficiency.

A multi-depot travelling salesman problem and its iterative and integrated approaches

This paper formulates a logistics distribution problem as the multi-depot travelling salesman problem (MDTSP). The decision makers not only have to determine the travelling sequence of the salesman for delivering finished products from a warehouse or depot to a customer, but also need to determine which depot stores which type of products so that the total travelling distance is minimised. The MDTSP is similar to the combination of the travelling salesman and quadratic assignment problems. In this paper, the two individual hard problems or models are formulated first. Then, the problems are integrated together, that is, the MDTSP. The MDTSP is constructed as both integer nonlinear and linear programming models. After formulating the models, we verify the integrated models using commercial packages, and most importantly, investigate whether an iterative approach, that is, solving the individual models repeatedly, can generate an optimal solution to the MDTSP. Copyright © 2006 Inderscience Enterprises Ltd.

Zero weights and non-zero slacks:different solution to the same problem

This paper re-assesses three independently developed approaches that are aimed at solving the problem of zero-weights or non-zero slacks in Data Envelopment Analysis (DEA). The methods are weights restricted, non-radial and extended facet DEA models. Weights restricted DEA models are dual to envelopment DEA models with restrictions on the dual variables (DEA weights) aimed at avoiding zero values for those weights; non-radial DEA models are envelopment models which avoid non-zero slacks in the input-output constraints. Finally, extended facet DEA models recognize that only projections on facets of full dimension correspond to well defined rates of substitution/transformation between all inputs/outputs which in turn correspond to non-zero weights in the multiplier version of the DEA model. We demonstrate how these methods are equivalent, not only in their aim but also in the solutions they yield. In addition, we show that the aforementioned methods modify the production frontier by extending existing facets or creating unobserved facets. Further we propose a new approach that uses weight restrictions to extend existing facets. This approach has some advantages in computational terms, because extended facet models normally make use of mixed integer programming models, which are computationally demanding.

Representation and parameterisation issues in genetic algorithms

A multi-chromosome GA (Multi-GA) was developed, based upon concepts from the natural world, allowing improved flexibility in a number of areas including representation, genetic operators, their parameter rates and real world multi-dimensional applications. A series of experiments were conducted, comparing the performance of the Multi-GA to a traditional GA on a number of recognised and increasingly complex test optimisation surfaces, with promising results. Further experiments demonstrated the Multi-GA's flexibility through the use of non-binary chromosome representations and its applicability to dynamic parameterisation. A number of alternative and new methods of dynamic parameterisation were investigated, in addition to a new non-binary 'Quotient crossover' mechanism. Finally, the Multi-GA was applied to two real world problems, demonstrating its ability to handle mixed type chromosomes within an individual, the limited use of a chromosome level fitness function, the introduction of new genetic operators for structural self-adaptation and its viability as a serious real world analysis tool. The first problem involved optimum placement of computers within a building, allowing the Multi-GA to use multiple chromosomes with different type representations and different operators in a single individual. The second problem, commonly associated with Geographical Information Systems (GIS), required a spatial analysis location of the optimum number and distribution of retail sites over two different population grids. In applying the Multi-GA, two new genetic operators (addition and deletion) were developed and explored, resulting in the definition of a mechanism for self-modification of genetic material within the Multi-GA structure and a study of this behaviour.

Bayesian analysis of the scatterometer wind retrieval inverse problem:Some new approaches

The retrieval of wind vectors from satellite scatterometer observations is a non-linear inverse problem. A common approach to solving inverse problems is to adopt a Bayesian framework and to infer the posterior distribution of the parameters of interest given the observations by using a likelihood model relating the observations to the parameters, and a prior distribution over the parameters. We show how Gaussian process priors can be used efficiently with a variety of likelihood models, using local forward (observation) models and direct inverse models for the scatterometer. We present an enhanced Markov chain Monte Carlo method to sample from the resulting multimodal posterior distribution. We go on to show how the computational complexity of the inference can be controlled by using a sparse, sequential Bayes algorithm for estimation with Gaussian processes. This helps to overcome the most serious barrier to the use of probabilistic, Gaussian process methods in remote sensing inverse problems, which is the prohibitively large size of the data sets. We contrast the sampling results with the approximations that are found by using the sparse, sequential Bayes algorithm.

A study of adaptive load balancing algorithms for distributed systems

With the advent of distributed computer systems with a largely transparent user interface, new questions have arisen regarding the management of such an environment by an operating system. One fertile area of research is that of load balancing, which attempts to improve system performance by redistributing the workload submitted to the system by the users. Early work in this field concentrated on static placement of computational objects to improve performance, given prior knowledge of process behaviour. More recently this has evolved into studying dynamic load balancing with process migration, thus allowing the system to adapt to varying loads. In this thesis, we describe a simulated system which facilitates experimentation with various load balancing algorithms. The system runs under UNIX and provides functions for user processes to communicate through software ports; processes reside on simulated homogeneous processors, connected by a user-specified topology, and a mechanism is included to allow migration of a process from one processor to another. We present the results of a study of adaptive load balancing algorithms, conducted using the aforementioned simulated system, under varying conditions; these results show the relative merits of different approaches to the load balancing problem, and we analyse the trade-offs between them. Following from this study, we present further novel modifications to suggested algorithms, and show their effects on system performance.

The development of constrained source localisation algorithms for human brain imaging

This work sets out to evaluate the potential benefits and pit-falls in using a priori information to help solve the Magnetoencephalographic (MEG) inverse problem. In chapter one the forward problem in MEG is introduced, together with a scheme that demonstrates how a priori information can be incorporated into the inverse problem. Chapter two contains a literature review of techniques currently used to solve the inverse problem. Emphasis is put on the kind of a priori information that is used by each of these techniques and the ease with which additional constraints can be applied. The formalism of the FOCUSS algorithm is shown to allow for the incorporation of a priori information in an insightful and straightforward manner. In chapter three it is described how anatomical constraints, in the form of a realistically shaped source space, can be extracted from a subject’s Magnetic Resonance Image (MRI). The use of such constraints relies on accurate co-registration of the MEG and MRI co-ordinate systems. Variations of the two main co-registration approaches, based on fiducial markers or on surface matching, are described and the accuracy and robustness of a surface matching algorithm is evaluated. Figures of merit introduced in chapter four are shown to given insight into the limitations of a typical measurement set-up and potential value of a priori information. It is shown in chapter five that constrained dipole fitting and FOCUSS outperform unconstrained dipole fitting when data with low SNR is used. However, the effect of errors in the constraints can reduce this advantage. Finally, it is demonstrated in chapter six that the results of different localisation techniques give corroborative evidence about the location and activation sequence of the human visual cortical areas underlying the first 125ms of the visual magnetic evoked response recorded with a whole head neuromagnetometer.

Exploiting uncertainty in nonlinear stochastic control problem

This work introduces a novel inversion-based neurocontroller for solving control problems involving uncertain nonlinear systems which could also compensate for multi-valued systems. The approach uses recent developments in neural networks, especially in the context of modelling statistical distributions, which are applied to forward and inverse plant models. Provided that certain conditions are met, an estimate of the intrinsic uncertainty for the outputs of neural networks can be obtained using the statistical properties of networks. More generally, multicomponent distributions can be modelled by the mixture density network. Based on importance sampling from these distributions a novel robust inverse control approach is obtained. This importance sampling provides a structured and principled approach to constrain the complexity of the search space for the ideal control law. The developed methodology circumvents the dynamic programming problem by using the predicted neural network uncertainty to localise the possible control solutions to consider. Convergence of the output error for the proposed control method is verified by using a Lyapunov function. Several simulation examples are provided to demonstrate the efficiency of the developed control method. The manner in which such a method is extended to nonlinear multi-variable systems with different delays between the input-output pairs is considered and demonstrated through simulation examples.

The effect of load on agent-based algorithms for distributed task allocation

Multi-agent algorithms inspired by the division of labour in social insects and by markets, are applied to a constrained problem of distributed task allocation. The efficiency (average number of tasks performed), the flexibility (ability to react to changes in the environment), and the sensitivity to load (ability to cope with differing demands) are investigated in both static and dynamic environments. A hybrid algorithm combining both approaches, is shown to exhibit improved efficiency and robustness. We employ nature inspired particle swarm optimisation to obtain optimised parameters for all algorithms in a range of representative environments. Although results are obtained for large population sizes to avoid finite size effects, the influence of population size on the performance is also analysed. From a theoretical point of view, we analyse the causes of efficiency loss, derive theoretical upper bounds for the efficiency, and compare these with the experimental results.