This paper studies a hybrid approach for the minimum conflict spanning tree (MCST) problem, where the MCST problem deals with finding a spanning tree (T) with the minimum number of conflicting edge-pairs. The problem ...
详细信息
ISBN:
(纸本)9783031539657;9783031539664
This paper studies a hybrid approach for the minimum conflict spanning tree (MCST) problem, where the MCST problem deals with finding a spanning tree (T) with the minimum number of conflicting edge-pairs. The problem finds some important real-world applications. In this hybrid approach (hSSGA), a steady-state genetic algorithm generates a child solution with the help of crossover operator and mutation operator which are applied in a mutually exclusive way, and the generated child solution is further improved through a local search based on reduction of conflicting edge-pairs. The proposed crossover operator is problem-specific operator that attempt to create a fitter child solution. All components of SSGA and local search effectively coordinate in finding a conflict-free solution or a solution with a minimal number of conflicting edge-pairs. Experimental results, particularly, on available 12 instances of type 1 benchmark instances whose conflict solutions are not known show that the proposed hybrid approach hSSGA is able to find better solution quality in comparison to state-of-the-art approaches. Also, hSSGA discovers new values on 8 instances out of 12 instances of type 1.
This paper presents an evolutionary algorithm-based steady-state grouping geneticalgorithm (SSGGA) for the single-machine scheduling problem with periodic machine availability (SinMSPMA problem) whose objective is to...
详细信息
Given a connected, undirected and edge-colored graph, the rainbow spanning forest (RSF) problem aims to find a rainbow spanning forest with the minimum number of rainbow trees, where a rainbow tree is a connected acyc...
详细信息
Given an undirected, connected and edge-weighted graph, the min-degree constrained minimum spanning tree (md-MST) problem aims to find a minimum spanning tree (T) in such a way that each non leaf vertex in T has degre...
详细信息
Given an undirected, connected and edge-weighted graph, the min-degree constrained minimum spanning tree (md-MST) problem aims to find a minimum spanning tree (T) in such a way that each non leaf vertex in T has degree at least d, where d is given a positive integer constant. This paper proposes a hybrid steady-state genetic algorithm (HSSGA) for the and-MST problem. The proposed HSSGA combines various components- such as problem-specific genetic operators (crossover operator and mutation operator) that are designed in such a way that the min-degree constraint is always maintained, hence resulting into a feasible child solution;and a population updating strategy that helps in maintaining diversity in the population- of the steady-state genetic algorithm with a fast local search. On a set of standard benchmark instances, the proposed HSSGA shows its effectiveness in comparison to state-ofthe-art approaches such as four versions of variable neighborhood search, branch and cut algorithm and three versions of generational geneticalgorithm. Moreover, HSSGA finds new best values on 32 out of 105 benchmark instances. (C) 2019 Elsevier B.V. All rights reserved.
This paper proposes a two-stage planning model for soft open point (SOP) and energy storage system (ESS) that considers the cost of faults in response to the current issue of SOP and ESS systems not considering the im...
详细信息
This paper proposes a two-stage planning model for soft open point (SOP) and energy storage system (ESS) that considers the cost of faults in response to the current issue of SOP and ESS systems not considering the impact of SOP access on load transfer in the event of a fault in the distribution network. Firstly, considering the uncertainty of "PV-load", typical scenarios of PV and load are constructed based on the clustering algorithm. Secondly, aiming at the economic performance of the distribution network and the capacity of PV access, a two-stage optimization model is established for the joint integration of SOP and ESS into the distribution network (normal and fault operation) under typical scenarios. The model is solved by using the second-order cone programming algorithm and steady-state genetic algorithm (SOCP-SSGA). Stage one involves planning for the integration capacity and location of SOP and ESS into the distribution network under each scenario within a period based on SOCP with the goal of minimizing economic costs. In stage two, the PV access capacity of the distribution network is optimized using SSGA with the goal of enhancing the PV accommodation capability. Finally, verification and analysis are conducted on an improved IEEE33 node system. The results show that when the system optimizes access to a group of SOP and ESS, the total economic cost is reduced by RMB 61,729 compared to random access, and the accessible PV capacity is increased by 0.5278 MW. Moreover, optimizing access to two sets of SOP and ESS can further reduce the total economic cost by RMB 107,048 compared to the optimized access group and increase accessible PV capacity by 1.5751 MW. Therefore, the proposed plan for SOP and ESS planning in this paper can significantly reduce the economic cost of distribution networks, enhance the absorption capacity of distributed photovoltaics, improve the voltage level of power grid operation, and, thereby, improve the economic and reliability of
The study of adversarial effects on AI systems is not a new concept, but much of the research has been devoted to deep learning. In this paper we explore the effects of adversarial examples on 4 machine learning class...
详细信息
ISBN:
(纸本)9781728183923
The study of adversarial effects on AI systems is not a new concept, but much of the research has been devoted to deep learning. In this paper we explore the effects of adversarial examples on 4 machine learning classifiers and measure the effectiveness of adversarial training. Additionally, we present a novel method for selecting adversarial training examples that lead to a more robust machine learning system. Our results suggest that adversarial examples can significantly hinder the classification performance and that adversarial training is an effective defensive counter-measure.
This paper introduces a new tool for context-sensitive grammar inference. The source code and library are publicly available via GitHub and NuGet repositories. The article describes the implemented algorithm, input pa...
详细信息
This paper introduces a new tool for context-sensitive grammar inference. The source code and library are publicly available via GitHub and NuGet repositories. The article describes the implemented algorithm, input parameters, and the produced output grammar. In addition, the paper contains several use-cases. The described library is written in F#, hence it can be used in any .NET Framework language (F#, C#, C++/CLI, Visual Basic, and J#) and run under the control of varied operating systems.
In this paper, we explore the effects of using evolutionary computation in an effort to increase the effectiveness of universal false positive and universal false negative attacks. These attacks are focused on machine...
详细信息
ISBN:
(纸本)9781728190488
In this paper, we explore the effects of using evolutionary computation in an effort to increase the effectiveness of universal false positive and universal false negative attacks. These attacks are focused on machine learning based fake news detection systems. For this task, we use a steady-state genetic algorithm to evolve a set of adversarial samples. Our results suggest that using a steady-state genetic algorithm to evolve adversarial samples yields a noticeable increase in misclassification rates and exposes potential vulnerabilities in the evaluated machine learning algorithms.
Given an undirected, connected, edge-weighted graph G and a positive integer d, the degree-constrained minimum spanning tree (dc-MST) problem aims to find a minimum spanning tree T on G subject to the constraint that ...
详细信息
Given an undirected, connected, edge-weighted graph G and a positive integer d, the degree-constrained minimum spanning tree (dc-MST) problem aims to find a minimum spanning tree T on G subject to the constraint that each vertex is either a leaf vertex or else has degree at most d in T, where d is a given positive integer. The dc-MST is NP-hard problem for d >= 2 and finds several real-world applications. This paper proposes a hybrid approach (HSSGA) combining a steady-state genetic algorithm and local search strategies for the this problem. An additional step (based on perturbation strategy at a regular interval of time) in the replacement strategy is applied in order to maintain diversity in the population throughout the search process. On a set of available 107 benchmark instances, computational results show the superiority of our proposed HSSGA in comparison with the state-of-the-art metaheuristic techniques.
Fake news is becoming an increasingly invasive problem within our society. As our society becomes more ingrained in technology, news has become more susceptible to technological predation. In this paper, we demonstrat...
详细信息
ISBN:
(纸本)9781728125473
Fake news is becoming an increasingly invasive problem within our society. As our society becomes more ingrained in technology, news has become more susceptible to technological predation. In this paper, we demonstrate how evolutionary-based feature selection increases fake news detection while dramatically reducing the number of features needed.
暂无评论