Benjamin N. Passow

Re-sampled inheritance search: high performance despite the simplicity

This paper proposes re-sampled inheritance search (RIS), a novel algorithm for solving continuous optimization problems. The proposed method, belonging to the class of Memetic Computing, is very simple and low demanding in terms of memory employment and computational overhead. The RIS algorithm is composed of a stochastic sample mechanism and a deterministic local search. The first operator randomly generates a solution and then recombines it with the best solution detected so far (inheritance) while the second operator searches in an exploitative way within the neighbourhood indicated by the stochastic operator. This extremely simple scheme is shown to display a very good performance on various problems, including hard to solve multi-modal, highly-conditioned, large scale problems. Experimental results show that the proposed RIS is a robust scheme that competitively performs with respect to recent complex algorithms representing the-state-of-the-art in modern continuous optimization. In order to further prove its applicability in real-world cases, RIS has been used to perform the control system tuning for yaw operations on a helicopter robot. Experimental results on this real-world problem confirm the value of the proposed approach.

Real-time evolution of an embedded controller for an autonomous helicopter

In this paper we evolve the parameters of a proportional, integral, and derivative (PID) controller for an unstable, complex and nonlinear system. The individuals of the applied genetic algorithm (GA) are evaluated on the actual system rather than on a simulation of it, thus avoiding the ldquoreality gaprdquo. This makes implicit a formal model identification for the implementation of a simulator. This also calls for the GA to be approached in an unusual way, where we need to consider new aspects not normally present in the usual situations using an unnaturally consistent simulator for fitness evaluation. Although elitism is used in the GAs, no monotonic increase in fitness is exhibited by the algorithm. Instead, we show that the GApsilas individuals converge towards more robust solutions.

Robustness analysis of evolutionary controller tuning using real systems

A genetic algorithm (GA) presents an excellent method for controller parameter tuning. In our work, we evolved the heading as well as the altitude controller for a small lightweight helicopter. We use the real flying robot to evaluate the GAs individuals rather than an artificially consistent simulator. By doing so we avoid the “reality gap”, taking the controller from the simulator to the real world. In this paper we analyze the evolutionary aspects of this technique and discuss the issues that need to be considered for it to perform well and result in robust controllers.

Dimensionality reduction and microarray data

Microarrays are being currently used for the expression levels of thousands of genes simultaneously. They present new analytical challenges because they have a very high input dimension and a very low sample size. It is highly complex to analyse multi-dimensional data with complex geometry and to identify lowdimensional “principal objects” that relate to the optimal projection while losing the least amount of information. Several methods have been proposed for dimensionality reduction of microarray data. Some of these methods include principal component analysis and principal manifolds. This article presents a comparison study of the performance of the linear principal component analysis and the non linear local tangent space alignment principal manifold methods on such a problem. Two microarray data sets will be used in this study. A classification model will be created using fully dimensional and dimensionality reduced data sets. To measure the amount of information lost with the two dimensionality reduction methods, the level of performance of each of the methods will be measured in terms of level of generalisation obtained by the classification models on previously unseen data sets. These results will be compared with the ones obtained using the fully dimensional data sets.

Real-world dynamic optimization using an adaptive-mutation compact genetic algorithm

While the interest in nature inspired optimization in dynamic environments has been increasing constantly over the past years, evaluations of some of these optimization algorithms are based on artificial benchmark problems. Little has been done to carry-out these evaluation using a real-world dynamic optimization problems. This paper presents a compact optimization algorithm for controllers in dynamic environments. The algorithm is evaluated using a real world dynamic optimization problem instead of an artificial benchmark problem, thus avoiding the reality gap. The experimental result shows that the algorithm has an impact on the performance of a controller in a dynamic environment. Furthermore, results suggest that evaluating the algorithms candidate solution using an actual real-world problem increases the controllers robustness.

Managing uncertainty in sound based control for an autonomous helicopter

In this paper we present our ongoing research using a multi-purpose, small and low cost autonomous helicopter platform (Flyper ). We are building on previously achieved stable control using evolutionary tuning. We propose a sound based supervised method to localise the indoor helicopter and extract meaningful information to enable the helicopter to further stabilise its flight and correct its flightpath. Due to the high amount of uncertainty in the data, we propose the use of fuzzy logic in the signal processing of the sound signature. We discuss the benefits and difficulties using type-1 and type-2 fuzzy logic in this real-time systems and give an overview of our proposed system.

Adapting traffic simulation for traffic management: A neural network approach

Static models and simulations are commonly used in urban traffic management but none feature a dynamic element for near real-time traffic control. This work presents an artificial neural network forecaster methodology applied to traffic flow condition prediction. The spatially distributed architecture uses life-long learning with a novel adaptive Artificial Neural Network based filter to detect and remove outliers from training data. The system has been designed to support traffic engineers in their decision making to react to traffic conditions before they get out of control. We performed experiments using feed-forward backpropagation, cascade-forward back-propagation, radial basis, and generalized regression Artificial Neural Networks for this purpose. Test results on actual data collected from the city of Leicester, UK, confirm our approach to deliver suitable forecasts.

Re-sampling search: A seriously simple memetic approach with a high performance

In the fashion of the Ockhams Razor principle for Memetic Computing approaches, this paper proposes an extremely simple and yet very efficient algorithm composed of two operators. The proposed approach employs a deterministic local search operator that periodically perturbs by means of a stochastic search component. The perturbation occurs by re-sampling the initial solution within the decision space. The deterministic local search is stopped by means of a precision based criterion and started over by means of the stochastic re-sampling. Although the concept of multi-start local search is not new in the optimization environment the proposed algorithm is shown to be extremely efficient on a broad set of diverse problems and competitive with complex algorithms representing the-state-of-the-art in computational intelligence optimization.

Meta-Heuristically Seeded Genetic Algorithm for Independent Job Scheduling in Grid Computing

Grid computing is an infrastructure which connects geographically distributed computers owned by various organizations allowing their resources, such as computational power and storage capabilities, to be shared, selected, and aggregated. Job scheduling problem is one of the most difficult tasks in grid computing systems. To solve this problem efficiently, new methods are required. In this paper, a seeded genetic algorithm is proposed which uses a meta-heuristic algorithm to generate its initial population. To evaluate the performance of the proposed method in terms of minimizing the makespan, the Expected Time to Compute (ETC) simulation model is used to carry out a number of experiments. The results show that the proposed algorithm performs better than other selected techniques.

Adaptive-mutation compact genetic algorithm for dynamic environments

In recent years, the interest in studying nature-inspired optimization algorithms for dynamic optimization problems (DOPs) has been increasing constantly due to its importance in real-world applications. Several techniques such as hyperselection, change prediction, hypermutation and many more have been developed to address DOPs. Among these techniques, the hypermutation scheme has proved beneficial for addressing DOPs, but requires that the mutation factors be picked a priori and this is one of the limitations of the hypermutation scheme. This paper investigates variants of the recently proposed adaptive-mutation compact genetic algorithm (amcGA). The amcGA is made up of a change detection scheme and mutation schemes, where the degree of change regulates the probability of mutation (i.e. the probability of mutation is directly proportional to the degree of change). This paper also presents a change trend scheme for the amcGA so as to boost its performance whenever a change occurs. Experimental results show that the change trend and mutation schemes have an impact on the performance of the amcGA in dynamic environment and also indicate that the effect of the schemes depends on the dynamics of the environment as well as the dynamic problem being considered.