Automatic parameter tuning for Evolutionary Algorithms using a Bayesian Case-Based Reasoning system

The widespread use and applicability of Evolutionary Algorithms is due in part to the ability to adapt them to a particular problem-solving context by tuning their parameters. This is one of the problems that a user faces when applying an Evolutionary Algorithm to solve a given problem. Before running the algorithm, the user typically has to specify values for a number of parameters, such as population size, selection rate, and probability operators.This paper empirically assesses the performance of an automatic parameter tuning system in order to avoid the problems of time requirements and the interaction of parameters. The system, based on Bayesian Networks and Case-Based Reasoning methodology, estimates the best parameter setting for maximizing the performance of Evolutionary Algorithms. The algorithms are applied to solve a basic problem in constraint-based, geometric parametric modeling, as an instance of general constraint-satisfaction problems.The experimental results demonstrate the validity of the proposed system and its potential effectiveness for configuring algorithms.     

Geometric optimization of yield-line patterns using a direct search method

A class of problems in the geometric optimization of yield-line patterns, for which the currently advocatedconjugate gradient andsequential linear programming geometric optimization algorithms fail is investigated. TheHooke-Jeeves direct search method is implemented and is demonstrated to solve such problems robustly.     

Stochastic system transformation using generalized orthonormal basis functions with applications to continuous-time system identification

This paper studies the system transformation using generalized orthonormal basis functions that include the Laguerre basis as a special case. The transformation of the deterministic systems is studied in the literature, which is called the Hambo transform. The aim of the paper is to develop a transformation theory for stochastic systems. The paper establishes the equivalence of continuous and transformed-discrete-time stochastic systems in terms of solutions. The method is applied to the continuous-time system identification problem. It is shown that using the transformed signals the PO-MOESP subspace identification algorithm yields consistent estimates for system matrices. An example is included to illustrate the efficacy of the proposed identification method, and to make a comparison with the method using the Laguerre filter.     

Sm@rtConfig: A context-aware runtime and tuning system using an aspect-oriented approach for data intensive engineering applications

Distributing the workload upon all available Processing Units (PUs) of a high-performance heterogeneous platform (e.g., PCs composed by CPU–GPUs) is a challenging task, since the execution cost of a task on distinct PUs is non-deterministic and affected by parameters not known a priori. This paper presents Sm@rtConfig, a context-aware runtime and tuning system based on a compromise between reducing the execution time of engineering applications and the cost of tasks' scheduling on CPU–GPUs' platforms. Using Model-Driven Engineering and Aspect Oriented Software Development, a high-level specification and implementation for Sm@rtConfig has been created, aiming at improving modularization and reuse in different applications. As case study, the simulation subsystem of a CFD application has been developed using the proposed approach. These system's tasks were designed considering only their functional concerns, whereas scheduling and other non-functional concerns are handled by Sm@rtConfig aspects, improving tasks modularity. Although Sm@rtConfig supports multiple PUs, in this case study, these tasks have been scheduled to execute on an platform composed by one CPU and one GPU. Experimental results show an overall performance gain of 21.77% in comparison to the static assignment of all tasks only to the GPU.     

Automatic parameter tuning with a Bayesian case-based reasoning system. A case of study

The parameter setting of an algorithm that will result in optimal performance differs across problem instance domains. Users spend a lot of time tuning algorithms for their specific problem domain, and this time could be saved by an automatic approach for parameter tuning.In this paper, we present a system that recommends the parameter configuration of an algorithm that solves a problem, conditioned by the particular features of the current problem instance to be solved. The proposed system is based on a basic adjustment model designed by authors (Pavon, R., Díaz, F., & Luzón, V. (2008). A model for parameter setting based on Bayesian networks. Engineering Applications of Artificial Intelligence, 21(1), 14–25) in which starting from experimental results concerning the search for solutions to several instances of the problem, a Bayesian network (BN) is induced and tries to infer the best configuration for the algorithm used.However, taking into account that the optimal parameter configuration may differ considerably across problem instances of a specific domain, the present work extends the former incorporating additional information about problem instances and using the case-based reasoning (CBR) methodology as the framework integrator for the different instances from the same problem, where each problem instance deals with a specific BN. In this way, the system will automatically recommend a parameter configuration for a given algorithm according to the characteristics of the problem instance at hand and past experience of similar instances.As an example, we empirically evaluate our Bayesian CBR system to tune a genetic algorithm for solving the root identification problem. The experimental results demonstrate the validity of the model proposed.     

Automatic tuning of communication protocols for vehicular ad hoc networks using metaheuristics

The emerging field of vehicular ad hoc networks (VANETs) deals with a set of communicating vehicles which are able to spontaneously interconnect without any pre-existing infrastructure. In such kind of networks, it is crucial to make an optimal configuration of the communication protocols previously to the final network deployment. This way, a human designer can obtain an optimal QoS of the network beforehand. The problem we consider in this work lies in configuring the File Transfer protocol Configuration (FTC) with the aim of optimizing the transmission time, the number of lost packets, and the amount of data transferred in realistic VANET scenarios. We face the FTC with five representative state-of-the-art optimization techniques and compare their performance. These algorithms are: Particle Swarm Optimization (PSO), Differential Evolution (DE), Genetic Algorithm (GA), Evolutionary Strategy (ES), and Simulated Annealing (SA). For our tests, two typical environment instances of VANETs for Urban and Highway scenarios have been defined. The experiments using ns- 2 (a well-known realistic VANET simulator) reveal that PSO outperforms all the compared algorithms for both studied VANET instances.     

Preference-based reinforcement learning: evolutionary direct policy search using a preference-based racing algorithm

We introduce a novel approach to preference-based reinforcement learning, namely a preference-based variant of a direct policy search method based on evolutionary optimization. The core of our approach is a preference-based racing algorithm that selects the best among a given set of candidate policies with high probability. To this end, the algorithm operates on a suitable ordinal preference structure and only uses pairwise comparisons between sample rollouts of the policies. Embedding the racing algorithm in a rank-based evolutionary search procedure, we show that approximations of the so-called Smith set of optimal policies can be produced with certain theoretical guarantees. Apart from a formal performance and complexity analysis, we present first experimental studies showing that our approach performs well in practice.     

Automatic tuning of pneumatic servo actuators

The high power-to-weight ratio and low cost of pneumatic actuators makes them attractive for robotics and automation applications; however, achieving fast, accurate position control with them is difficult. Furthermore, tuning the controller (and/or obtaining process models) requires extensive expertise and time. In this paper, a novel automatic tuning methodology for the accurate position control of pneumatic servo actuators is described. The methodology combines off-line model-based analysis with on-line iteration. Experiments performed on three actuators with distinct open-loop dynamics verified the auto-tuner's effectiveness. The auto-tuning required 1/50 of the time needed for manual tuning and only non-expert supervision. The performance of the auto-tuned actuators is comparable to that achieved by other researchers using more complex and expensive hardware.     

Ecological applications using a novel expert system shell

Much of the information used by ecologists in modelling and decision making is imprecise. The imprecision arises both from data that are inexact or incomplete and from the use of ecological principles that are sometimes less than fully reliable and may be conflicting. Nevertheless, expert ecologists are able to construct usable models and make decisions that are used to manage and control ecological resources. This paper describes a unique expert system shell, developed in conjunction with user ecologists, which incorporates features enabling ecologists to represent knowledge and uncertainty in their expert systems in a way that is natural and appropriate. The reasoning mechanism was similarly developed in conjunction with user ecologists. It produces solutions to a class of expert level problems along with explanatory mechanisms and an appropriate analysis of the reasoning process. Three expert systems have been constructed by ecologists using this expert system shell. This enabled the shell designers to evaluate features for inclusion in the shell. The successful use of the shell by the ecologists has shown that significant economies arise when expert system shell design is tailored to use by a specific class of experts, in this case ecologists.     

Automatic folding of cartons using a reconfigurable robotic system

This paper proposes new theories and methodology for the study of folding carton and reconfigurable packaging systems and presents both virtual and experimental systems to support the theories. Equivalent mechanisms of cartons were established by describing carton creases as joints and carton panels as links. With the analysis of the equivalent mechanism, the gusset vertexes of cartons were analyzed based on their equivalent spherical linkages and were identified as guiding linkages that determine folding. A reconfigurable robotic system was developed to demonstrate the ability to erect diverted cartons in the reconfigurable packaging system which could previously only be achieved manually.     

Automatic hexahedral mesh generation using a new grid-based method with geometry and mesh transformation

A geometry and mesh transformation approach is proposed to overcome the traditional problem of poorly shaped elements at the boundary using the grid-based method of mesh generation. This is achieved by transforming the original geometric model to a topologically similar recognition model which conforms fully to the Cartesian directions. Such a recognition model is constructed by tessellating the original model and then employing a fuzzy logic method to determine the normal directions of the faces. A three-dimensional field morphing algorithm is used to position the features of the recognition model. Such a recognition model is then meshed with hexahedral elements without any degeneracy using the new grid-based algorithm. The mesh of the recognition model is mapped back to the original geometric model by employing a transformation based on the Laplacian-isoparametric equation.     

Automatic end-of-line tuning for a motion inverter in agricultural tractors

End-of-line tuning is a crucial step for any mass-produced system endowed with automatic controllers. As a matter of fact, due to components tolerances and spreads in the production line, the controller tuning performed on a prototype system is never optimal on the final product. In many industrial applications, though, the end-of-line tuning is performed by human testers, and this does not always guarantee an objective assessment of the closed-loop system quality. This paper proposes a systematic way to design an automatic tuning procedure for a motion-inverter controller in agricultural tractors, which allows to significantly reduce the costs of end-of-line tuning and to obtain a homogeneous manoeuvre quality in all vehicles. The proposed automatic tuning system adapts the controller parameters governing the open-loop phase of the manoeuvre until a predefined manoeuvre quality is achieved. The parameters adaptation phase is guided by an on-line objective assessment of the manoeuvre quality from measured data, which allows to automatically classify the performed manoeuvre with respect to its quality attributes. The effectiveness of the proposed approach is assessed on a prototype vehicle.     

Controller gain tuning of a simultaneous multi-axis PID control system using the Taguchi method

This paper presents a gain-tuning scheme for multi-axis PID control systems using the Taguchi method. A parallel-mechanism machine tool has been selected as an experimental set-up. This machine has eight servodrivers and each servodriver has four controller gains, resulting in a total of 32 controller gains to be tuned. Through a series of `Design of Experiments’ suggested by the Taguchi method, an optimal and robust set of PID controller gains has been obtained. The index of aggregate position and velocity errors has been reduced to 61.4%, regardless of feedrate variation, after the experimental gain tuning.     

Online tuning of a supervisory fuzzy controller for low-energy building system using reinforcement learning

This paper proposes a model-free method using reinforcement learning scheme to tune a supervisory controller for a low-energy building system online. The training time and computational demands are reduced by basing the supervisor on sets of fuzzy rules generated by off-line optimisation and by learning the optimal values of only one parameter, which selects the most appropriate set of rules. By carefully choosing the tuning targets, discretizing the state space, parameterizing the fuzzy rule base, using fuzzy trace-back, the proposed method can complete the training process in one season.     

A modified exhaustive search on a password system using SHA-1

In this paper, we show a method of exhaustive search on a password system that uses SHA-1 iteratively. Our method uses both the technique shown in Steube [16] and a technique for computing repetitions of SHA-1. Combining these two techniques reduces the total number of operations. We also show how to apply our method to MS Office (Microsoft Office) 2007/2010.     

Source transformation in software engineering using the TXL transformation system

Many tasks in software engineering can be characterized as source to source transformations. Design recovery, software restructuring, forward engineering, language translation, platform migration, and code reuse can all be understood as transformations from one source text to another. The tree transformation language, TXL, is a programming language and rapid prototyping system specifically designed to support rule-based source to source transformation. Originally conceived as a tool for exploring programming language dialects, TXL has evolved into a general purpose source transformation system that has proven well suited to a wide range of software maintenance and reengineering tasks, including the design recovery, analysis and automated reprogramming of billions of lines of commercial Cobol, PL/I, and RPG code for the Year 2000. In this paper, we introduce the basic features of modern TXL and its use in a range of software engineering applications, with an emphasis on how each task can be achieved by source transformation.     

Robust control system design using random search and geneticalgorithms

Random search and genetic algorithms find compensators to minimize stochastic robustness cost functions. Statistical tools are incorporated in the algorithms, allowing intelligent decisions to be based on “noisy” Monte Carlo estimates. The genetic algorithm includes clustering analysis to improve performance and is significantly better than the random search for this application. The algorithm is used to design a compensator for a benchmark problem, producing a control law with excellent stability and performance robustness     

基于蓝芽通讯的自动抄表系统

介绍一种使用蓝芽无线技术(Bluetooth)和现场可编程门数组(FPGA)数据控制的智能遥测读表系统,能自动量度仪表的资料并实时转送到计算机作数字辨识处理。该辨识系统是通过ANN仿脑神经网络而设计;论述了整个系统结构、运作原理、硬件设计及其测试结果。