Multicriteria optimization based on Pareto dominance becomes ineffective as the number of objectives increases. We analyze an alternative offered by Lorenz dominance for multiobjective optimization. Our aim is to stud...
详细信息
ISBN:
(纸本)9780769549347;9781467350266
Multicriteria optimization based on Pareto dominance becomes ineffective as the number of objectives increases. We analyze an alternative offered by Lorenz dominance for multiobjective optimization. Our aim is to study whether Lorenz dominance improves the scalability of evolutionary techniques. The set of Lorenz-optimal solutions is a subset of Pareto-optimal solutions. Experiments indicate that the Lorenz-optimal set contains only those optimal solutions that equally optimize all criteria. evolutionary optimization algorithms based on Lorenz dominance usually are able to detect optimal solutions even when the number of objectives is large (m > 3).
evolutionary computation (EC) is a growing research field of Artificial Intelligence (AI), particularly of computational Intelligence (CI). EC is the general term for several computational techniques which use ideas a...
详细信息
ISBN:
(纸本)9789604740994
evolutionary computation (EC) is a growing research field of Artificial Intelligence (AI), particularly of computational Intelligence (CI). EC is the general term for several computational techniques which use ideas and get inspiration from natural evolution/adaptation and is divided in two main areas: the evolutionary Algorithms (EA) and the Swarm Intelligence (SI). This paper presents hybridization between an EA algorithm - the Genetic Algorithm (GA) and a Si algorithm - the Particle Swarm Optimization Algorithm (PSO). The resulting algorithm is applied to the synthesis of combinational logic circuits. With this combination is possible to take advantage of the best features of each particular algorithm.
The design of a spacecraft is an evolutionary process that starts from requirements and evolves over time across different design phases. During this process, a lot of changes can happen. They can affect mass and powe...
详细信息
ISBN:
(纸本)9781467318112
The design of a spacecraft is an evolutionary process that starts from requirements and evolves over time across different design phases. During this process, a lot of changes can happen. They can affect mass and power at component level, at subsystem level, and even at system level. Each spacecraft has to meet overall constraints in terms of mass and power: for this reason, it's important to be sure that the design does not exceed these limitations. Current practice in system modeling deals with this problem by allocating margins on single components and on each of the subsystems. However, a statistical characterization of these fluctuations in mass and power is missing, and the consequence is a design that either is too risky and does not fit the mission constraints, or is too conservative and generates an inefficient utilization of resources. Hence, the objective of this research is to develop a mathematical approach to quantify the likelihood that the design would meet the spacecraft and mission constraints while the design matures. Due to the complexity of the problem and to the different expertise and knowledge required to develop a complete risk model for all the different subsystems, the research is focused on risk estimation for a specific subsystem: communication. Communication constitutes a key design driver in many different spacecraft, and it is also the core in the design of commercial satellite applications. Moreover, the current research aims to be a "proof of concept," which can then be further expanded to the different subsystems, as well as to the whole spacecraft design process. Particularly important in this analysis is the development of optimization frameworks to compare different design architectures, and to select the one that achieves design objectives, like minimal mass and power consumption, while minimizing the risk associated with these same metrics. The article is structured as follows: an overview of the model to perform statistical r
This paper proposes an Artificial Neural Network (ANN) based daily peak load forecasting method by differential evolutionary particle swarm optimization (DEEPSO) considering outliers. When outliers exist in the traini...
详细信息
This paper proposes an Artificial Neural Network (ANN) based daily peak load forecasting method by differential evolutionary particle swarm optimization (DEEPSO) considering outliers. When outliers exist in the training data, forecasting accuracy of daily peak load forecasting can be affected by the outliers. Therefore, engineers have removed the outliers from training data so far and it is a heavy burden for engineers. Utilization of evolutionary computation has a possibility to solve this problem. Moreover, forecasting accuracy may be improved using evolutionary computation techniques instead of the conventional stochastic gradient descent (SGD) with outliers. The proposed weights tuning method by DEEPSO is compared with the conventional weights tuning methods by SGD and PSO for verification of the efficacy of the proposed method. (C) 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Analog design is a bottleneck in the design of integrated circuits. A recently proposed method to cope with the complexity of analog design is the use of a multi-objective bottom-up flow, which makes use of the concep...
详细信息
ISBN:
(纸本)9781424450435
Analog design is a bottleneck in the design of integrated circuits. A recently proposed method to cope with the complexity of analog design is the use of a multi-objective bottom-up flow, which makes use of the concept of Pareto-optimal front (POF) to capture performance trade-offs of analog components, and through which these can be exploited during top-down design of a complex (hierarchically-wise) analog circuit. In this paper, we describe a step forward and transform this technique, through a new type of front we call Multi-Mode Pareto-optimal Front, to design reconfigurable Analog-to-Digital Converters (ADCs). We demonstrate that not only design time is shortened but also that design complexity of reconfigurable circuits can be more systematically and efficiently managed.
This paper discusses measurement methods of human behaviors based on sensor network and human interaction of rehabilitation using robot partners. First, we explain robot partners and sensor networks for rehabilitation...
详细信息
ISBN:
(纸本)9781424473175
This paper discusses measurement methods of human behaviors based on sensor network and human interaction of rehabilitation using robot partners. First, we explain robot partners and sensor networks for rehabilitation. Next, we apply a steady-state genetic algorithm to extract human motions from 3D distance image. Finally, we discuss the effectiveness of the proposed methods through several experimental results.
This paper investigates the evolution of group tactics and counter tactics for wargaming and real-time strategy games. Inspired by potential field methods in robotics, we compactly represent group behavior as a combin...
详细信息
ISBN:
(纸本)9781467359115
This paper investigates the evolution of group tactics and counter tactics for wargaming and real-time strategy games. Inspired by potential field methods in robotics, we compactly represent group behavior as a combination of several potential fields and evolve potential field parameters against hand-coded opponent groups. A novel real-coded evolutionary algorithm encourages tactic diversity by using a new diversity metric to mediate parent selection for recombination. Preliminary results indicate that we can quickly evolve counter tactics that beat hard coded opponent groups.
We present a model-agnostic framework for jointly optimizing the predictive performance and interpretability of supervised machine learning models for tabular data. Interpretability is quantified via three measures: f...
详细信息
ISBN:
(纸本)9798400701191
We present a model-agnostic framework for jointly optimizing the predictive performance and interpretability of supervised machine learning models for tabular data. Interpretability is quantified via three measures: feature sparsity, interaction sparsity of features, and sparsity of non-monotone feature effects. By treating hyperparameter optimization of a machine learning algorithm as a multi-objective optimization problem, our framework allows for generating diverse models that trade off high performance and ease of interpretability in a single optimization run. Efficient optimization is achieved via augmentation of the search space of the learning algorithm by incorporating feature selection, interaction and monotonicity constraints into the hyperparameter search space. We demonstrate that the optimization problem effectively translates to finding the Pareto optimal set of groups of selected features that are allowed to interact in a model, along with finding their optimal monotonicity constraints and optimal hyperparameters of the learning algorithm itself. We then introduce a novel evolutionary algorithm that can operate efficiently on this augmented search space. In benchmark experiments, we show that our framework is capable of finding diverse models that are highly competitive or outperform state-of-the-art XGBoost or Explainable Boosting Machine models, both with respect to performance and interpretability.
Genetic programming (GP) has been shown to be very effective for performing data mining tasks. Despite this, it has seen relatively little use in clustering. In this work, we introduce a new GP approach for performing...
详细信息
ISBN:
(纸本)9781450349208
Genetic programming (GP) has been shown to be very effective for performing data mining tasks. Despite this, it has seen relatively little use in clustering. In this work, we introduce a new GP approach for performing graph-based (GPGC) non-hyper-spherical clustering where the number of clusters is not required to be set in advance. The proposed GPGC approach is compared with a number of well known methods on a large number of data sets with a wide variety of shapes and sizes. Our results show that GPGC is the most generalisable of the tested methods, achieving good performance across all datasets. GPGC significantly outperforms all existing methods on the hardest ellipsoidal datasets, without needing the user to pre-define the number of clusters. To our knowledge, this is the first work which proposes using GP for graph-based clustering.
We examine the use of an evolutionary algorithm to design a feedback controller for a dual pursuit-evasion problem. In this problem, two players, Player A and Player B, move about an obstacle-free, two-dimensional pla...
详细信息
ISBN:
(纸本)9781509046010
We examine the use of an evolutionary algorithm to design a feedback controller for a dual pursuit-evasion problem. In this problem, two players, Player A and Player B, move about an obstacle-free, two-dimensional plane with constant speeds and bounded turn rates. Player A strives to capture Player B by maneuvering behind and closing within a defined capture distance. Simultaneously, Player B is attempting to capture Player A while avoiding being captured itself. Although the general form of this problem is two-sided, we examine the design of strategies for Player A against a collection of possible adversarial strategies implemented by Player B. We pose a nearest neighbor switching control structure that is represented using a parameterized matrix. An evolutionary algorithm is utilized to evolve these parameters in order to develop a feedback controller for Player A to efficiently capture Player B while evading capture itself.
暂无评论