Configuring an evolutionary Algorithm (EA) can be a haphazard and inefficient process. An EA practitioner may have to choose between a plethora of search operator types and other parameter settings. In contrast, the g...
详细信息
ISBN:
(纸本)9781509042401
Configuring an evolutionary Algorithm (EA) can be a haphazard and inefficient process. An EA practitioner may have to choose between a plethora of search operator types and other parameter settings. In contrast, the goal of EA principled design is a more streamlined and systematic design methodology, which first seeks to better understand the problem domain, and only then uses such acquired insights to guide the choice of parameters and operators. We introduce a new approach to principled design of EAs based on kernel methods. Several popular machine learning and data analysis paradigms, which have been successfully applied to a wide range of difficult real world problems, would fall under this kernel umbrella. We demonstrate how kernel functions, which capture useful problem domain knowledge, can be used to directly construct EA search operators. We test two kernel search operators on a suite of four challenging combinatorial optimization problem domains. These novel kernel search operators exhibit superior performance to some traditional EA search operators.
We consider the task of determining the pose of a depth camera based on a single target depth image and a 3D model of the indoor environment that the image was taken in. We identify the quality of a pose estimate with...
详细信息
ISBN:
(纸本)9781509006229
We consider the task of determining the pose of a depth camera based on a single target depth image and a 3D model of the indoor environment that the image was taken in. We identify the quality of a pose estimate with summed differences between depth values in the target depth image and a depth image generated synthetically by using that pose estimate in the 3D model. We then propose an evolutionary algorithm for optimizing pose estimates. The performance of that algorithm is evaluated in two artificial test environments, and perspectives for use of the algorithm in real environments are discussed.
Efficient usage of the knowledge provided by the Linked Data community is often hindered by the need for domain experts to formulate the right SPARQL queries to answer questions. For new questions they have to decide ...
详细信息
ISBN:
(纸本)9783319490045;9783319490038
Efficient usage of the knowledge provided by the Linked Data community is often hindered by the need for domain experts to formulate the right SPARQL queries to answer questions. For new questions they have to decide which datasets are suitable and in which terminology and modelling style to phrase the SPARQL query. In this work we present an evolutionary algorithm to help with this challenging task. Given a training list of source-target node-pair examples our algorithm can learn patterns (SPARQL queries) from a SPARQL endpoint. The learned patterns can be visualised to form the basis for further investigation, or they can be used to predict target nodes for new source nodes. Amongst others, we apply our algorithm to a dataset of several hundred human associations (such as "circle-square") to find patterns for them in DBpedia. We show the scalability of the algorithm by running it against a SPARQL endpoint loaded with > 7.9 billion triples. Further, we use the resulting SPARQL queries to mimic human associations with a Mean Average Precision (MAP) of 39.9% and a Recall@10 of 63.9%.
Most real world optimization problems involve constraints and constraint handling has long been an area of active research. While older techniques explicitly preferred feasible solutions over infeasible ones, recent s...
详细信息
ISBN:
(纸本)9783319282701;9783319282695
Most real world optimization problems involve constraints and constraint handling has long been an area of active research. While older techniques explicitly preferred feasible solutions over infeasible ones, recent studies have uncovered some shortcomings of such strategies. There has been a growing interest in the efficient use of infeasible solutions during the course of search and this paper presents of short review of such techniques. These techniques prefer good infeasible solutions over feasible solutions during the course of search (or a part of it). The review looks at major reported works over the years and outlines how these preferences have been dealt in various stages of the solution process, viz, problem formulation, parent selection/recombination and ranking/selection. A tabular summary is then presented for easy reference to the work in this area.
Complex Event Processing (CEP) is an emerging technology to process streaming data and to generate response actions in real time. CEP systems treat all sensor data as primitive events and attempt to detect semanticall...
详细信息
ISBN:
(纸本)9780769556703
Complex Event Processing (CEP) is an emerging technology to process streaming data and to generate response actions in real time. CEP systems treat all sensor data as primitive events and attempt to detect semantically high level events and related actions by matching them using event patterns. These event patterns are the rules which combine primitive events according to temporal, logical, or spatial correlations among them. Although event patterns (decision rules) can be provided by experts in simplistic scenarios, the huge amount of sensor data makes this unfeasible. The main purpose of the underlying paper is replacing manual identification of event patterns. Considering the uncertainty related to the sensor data, Fuzzy Unordered Rule Induction Algorithm (FURIA) was implemented to identify event patterns after selecting the relevant feature subset using Elitist Pareto-based Multi Objective evolutionary Algorithm for Diversity Reinforcement (ENORA). The results were compared to the alternative machine learning approaches.
Nowadays, algorithmic studies of multi-objective evolutionary algorithms (MOEAs) are flooded with too many search algorithms. Each MOEA has its own expert problem domain. To clarify not only the optimal MOEA and its p...
详细信息
ISBN:
(纸本)9781509026784
Nowadays, algorithmic studies of multi-objective evolutionary algorithms (MOEAs) are flooded with too many search algorithms. Each MOEA has its own expert problem domain. To clarify not only the optimal MOEA and its parameters for each of multiple multi-objective optimization problems (MOPs) but the robust MOEAs for multiple MOPs, this work proposes a meta-MOEA framework to search the Pareto optimal algorithmic parameters for multiple MOPs. In this work, we use two DTLZ2 benchmark problems with 2 and 4 objectives and optimize the base algorithm, the crossover rate and its parameter, the mutation rate and its parameter for the both DTLZ2 problems by the meta-MOEA. The experiment results show that the optimal algorithmic parameters for each of two DTLZ2 problems are different and the robust algorithmic parameters for both problems can be obtained by the meta-MOEA framework.
The main objective of discretization is to transform numerical attributes into discrete ones. The intention is to provide the possibility to use some learning algorithms which require discrete data as input and to hel...
详细信息
ISBN:
(纸本)9783319262277;9783319262253
The main objective of discretization is to transform numerical attributes into discrete ones. The intention is to provide the possibility to use some learning algorithms which require discrete data as input and to help the experts to understand the data more easily. Due to the fact that in classification problems there are high interactions among multiple attributes, we propose the use of evolutionary algorithms to select a subset of cut points for multivariate discretization based on a wrapper fitness function. The algorithm proposed has been compared with the best state-of-the-art discretizers with two decision trees-based classifiers: C4.5 and PUBLIC. The results reported indicate that our proposal outperforms the rest of the discretizers in terms of accuracy and requiring a lower number of intervals.
Paper presents the application of evolutionary algorithms and polynomial interpolation in ship evolutionary trajectory planning method and its comparison to classic approach, where trajectory is modeled by straight li...
详细信息
ISBN:
(纸本)9783319231808;9783319231792
Paper presents the application of evolutionary algorithms and polynomial interpolation in ship evolutionary trajectory planning method and its comparison to classic approach, where trajectory is modeled by straight lines. evolutionary algorithms are group of methods that allows\to find a collision free trajectory in real time, while polynomial interpolation allows to model smooth trajectory, which keeps continuity of velocity and acceleration values along path in opposition to straight lines approach. Paper presents the experimental researches for several collision situations at sea with application of trajectory modeled by straight lines and polynomial interpolation.
In order to effectively design nearly Zero Energy Buildings, the assessment of energy performance in the early design stages through simulation is an important, although very demanding and complex, procedure. Over the...
详细信息
In order to effectively design nearly Zero Energy Buildings, the assessment of energy performance in the early design stages through simulation is an important, although very demanding and complex, procedure. Over the last decades, various tools and methods have been developed to address performance-related design questions, mostly using Multi-Objective Optimization algorithms. Technological advances have revolutionized the way Architects design and think, automating complex tasks and allowing the assessment of multiple variants at the same time. In this paper, a new nZEB design workflow methodology is proposed, integrating evolutionary algorithms and energy simulation, and its capabilities and current limitations are explored.
This paper investigates the use of a metaheuristic evolutionary algorithm. The algorithm, known as Bird Mating Optimizer (BMO), allows fault diagnosis and state-of-health estimation by comparing the battery parameters...
详细信息
ISBN:
(纸本)9781509009169
This paper investigates the use of a metaheuristic evolutionary algorithm. The algorithm, known as Bird Mating Optimizer (BMO), allows fault diagnosis and state-of-health estimation by comparing the battery parameters with the values for a brand new battery, in real time, which in turn allows the battery management system (BMS) to improve the energy management and eventually the lifetime of the battery. In this study, the equivalent-circuit model (ECM) parameters of a lead-acid battery are extracted from its voltage response using the BMO algorithm. The accuracy of the BMO method is then compared to traditional Least Square (LS) algorithm. The BMO extracted parameters showed close fit to the experimental data.
暂无评论