Оглавление
Patrick Siarry. Optimization and Machine Learning
Table of Contents
List of Illustrations
List of Tables
Guide
Pages
Optimization and Machine Learning. Optimization for Machine Learning and Machine Learning for Optimization
Introduction
1. Vehicle Routing Problems with Loading Constraints: An Overview of Variants and Solution Methods
1.1. Introduction
1.2. The capacitated vehicle routing problem with two-dimensional loading constraints
1.2.1. Solution methods
1.2.2. Problem description
1.2.3. The 2L-CVRP variants
1.2.3.1. The 2L-CVRP with time constraints
1.2.3.2. The 2L-CVRP with backhaul
1.2.3.3. 2L-CVRP with pickup and delivery constraints
1.2.4. Computational analysis
1.3. The capacitated vehicle routing problem with three-dimensional loading constraints
1.3.1. Solution methods
1.3.2. Problem description
1.3.3. 3L-CVRP variants
1.3.3.1. 3L-CVRP with time windows
1.3.3.2. 3L-CVRP with backhaul
1.3.3.3. 3L-CVRP with pickup and delivery
1.3.3.4. 3L-CVRP with split delivery
1.3.4. Computational analysis
1.4. Perspectives on future research
1.5. References
2. MAS-aware Approach for QoS-based IoT Workflow Scheduling in Fog-Cloud Computing
2.1. Introduction
2.2. Related works
2.3. Problem formulation
2.3.1. IoT-workflow modeling
2.3.2. Resources modeling
2.3.3. QoS-based workflow scheduling modeling
2.4. MAS-GA-based approach for IoT workflow scheduling
2.4.1. Architecture model
2.4.2. Multi-agent system model
2.4.3. MAS-based workflow scheduling process
Algorithm 2.1: MAS-GA-based algorithm
2.5. GA-based workflow scheduling plan
2.5.1. Solution encoding
2.5.2. Fitness function
2.5.2.1. Crossover operator
2.5.3. Mutation operator
2.6. Experimental study and analysis of the results
2.6.1. Experimental results
2.7. Conclusion
2.8. References
3. Solving Feature Selection Problems Built on Population-based Metaheuristic Algorithms
3.1. Introduction
3.2. Algorithm inspiration
3.2.1. Wolf pack hierarchy
3.2.1.1. Alpha wolf
3.2.1.2. Beta wolf
3.2.1.3. Delta wolf
3.2.1.4. Omega wolf
3.2.2. The four phases of pack hunting
3.3. Mathematical modeling. 3.3.1. Pack hierarchy
3.3.2. Four phases of hunt modeling. 3.3.2.1. Search and pursue
3.3.2.2. Encirclement and harassment
3.3.2.3. Attack
3.3.3. Research phase – exploration
3.3.4. Attack phase – exploitation
3.3.5. Grey wolf optimization algorithm pseudocode. Algorithm 3.1. GWO pseudo code
3.4. Theoretical fundamentals of feature selection
3.4.1. Feature selection definition
3.4.2. Feature selection methods
3.4.3. Filter method
3.4.4. Wrapper method
3.4.5. Binary feature selection movement
3.4.5.1. Forward selection
3.4.5.2. Backward elimination
3.4.6. Benefits of feature selection for machine learning classification algorithms
3.5. Mathematical modeling of the feature selection optimization problem
3.5.1. Optimization problem definition
3.5.2. Binary discrete search space
3.5.3. Objective functions for the feature selection
3.5.3.1. Accuracy
3.5.3.2. Error rate
3.5.3.3. True positive (TP)
3.5.3.4. True negative (TN)
3.5.3.5. False positive (FP)
3.5.3.6. False negative (FN)
3.6. Adaptation of metaheuristics for optimization in a binary search space
3.6.1. Module M1
3.6.2. Module M2
3.6.2.1. Change rate rule R1
3.6.2.2. Standard rule
3.6.2.3. 1’s complement rule R3
3.6.2.4. Statistic rule R4
3.7. Adaptation of the grey wolf algorithm to feature selection in a binary search space
3.7.1. First algorithm bGWO1
3.7.1.1. Module M1
3.7.1.2. Module M2
3.7.2. Second algorithm bGWO2
3.7.3. Algorithm 2: first approach of the binary GWO
3.7.4. Algorithm 3: second approach of the binary GWO
3.8. Experimental implementation of bGWO1 and bGWO2 and discussion
3.9. Conclusion
3.10. References
4. Solving the Mixed-model Assembly Line Balancing Problem by using a Hybrid Reactive Greedy Randomized Adaptive Search Procedure
4.1. Introduction
4.2. Related works from the literature
4.3. Problem description and mathematical formulation. 4.3.1. Problem description
4.3.2. Mathematical formulation
4.4. Basic greedy randomized adaptive search procedure
4.5. Reactive greedy randomized adaptive search procedure
4.6. Hybrid reactive greedy randomized adaptive search procedure for the mixed model assembly line balancing problem type-2
4.6.1. The proposed construction phase
4.6.2. The local search phase
4.7. Experimental examples
4.7.1. Results and discussion
4.8. Conclusion
4.9. References
5. An Interactive Attention Network with Stacked Ensemble Machine Learning Models for Recommendations
5.1. Introduction
5.2. Related work. 5.2.1. Attention network mechanism in recommender systems
5.2.2. Stacked machine learning for optimization
5.3. Interactive personalized recommender
5.3.1. Notation
5.3.2. The interactive attention network recommender
Algorithm 1: The Interactive Attention network recommender
5.3.3. The stacked content-based filtering recommender
5.4. Experimental settings
5.4.1. The datasets
5.4.2. Evaluation metrics
5.4.3. Baselines
5.5. Experiments and discussion. 5.5.1. Hyperparameter analysis
5.5.2. Performance comparison with the baselines
5.6. Conclusion
5.7. References
6. A Comparison of Machine Learning and Deep Learning Models with Advanced Word Embeddings: The Case of Internal Audit Reports
6.1. Introduction
6.2. Related work
6.2.1. Word embedding
6.2.2. Deep learning models. 6.2.2.1. CNN architecture
6.2.2.2. BERT architecture
6.2.2.3. LSTM architecture
6.3. Experiments and evaluation
6.4. Conclusion and future work
6.5. References
7. Hybrid Approach based on Multi-agent System and Fuzzy Logic for Mobile Robot Autonomous Navigation
7.1. Introduction
7.2. Related works
7.2.1. Classical approaches
7.2.1.1. Methods based on the configuration of space
7.2.1.2. Potential field methods
7.2.2. Advanced methods
7.2.2.1. Navigation with fuzzy logic
7.2.2.2. Artificial Neuronal Network
7.2.2.3. Metaheuristic
7.2.2.4. Multi-agent systems
7.3. Problem position
7.4. Developed control architecture
7.4.1. Agents description. 7.4.1.1. Perception agent
7.4.1.2. Locomotion agent
7.4.1.3. Feasibility agent
7.4.1.4. Fuzzy controller agent
7.5. Navigation principle by fuzzy logic. 7.5.1. Fuzzy logic overview
7.5.2. Description of simulated robot
7.5.3. Strategy of navigation
7.5.4. Fuzzy controller agent
7.6. Simulation and results
7.7. Conclusion
7.8. References
8. Intrusion Detection with Neural Networks: A Tutorial
8.1. Introduction. 8.1.1. Intrusion detection systems
8.1.2. Artificial neural networks
8.1.3. The NSL-KDD dataset
8.2. Dataset analysis
8.2.1. Dataset summary
8.2.2. Features
8.2.3. Binary feature distribution
8.2.4. Categorical features distribution
8.2.5. Numerical data distribution
8.2.6. Correlation matrix
8.3. Data preparation
8.3.1. Data cleaning
8.3.2. Categorical columns encoding
8.3.3. Normalization
8.4. Feature selection
8.4.1. Tree-based selection
8.4.2. Univariate selection
8.5. Model design. 8.5.1. Project environment
8.5.2. Building the neural network
8.5.3. Learning hyperparameters
8.5.4. Epochs
8.5.5. Batch size
8.5.6. Dropout layers
8.5.7. Activation functions
8.6. Results comparison
8.6.1. Evaluation metrics
8.6.2. Preliminary models
8.6.2.1. Shallow network with “relu” activation
8.6.2.2. Shallow network with “tanh” and “sigmoid” activation
8.6.2.3. Conclusions
8.6.3. Adding dropout
8.6.4. Adding more layers
8.6.5. Adding feature selection
8.7. Deployment in a network
8.7.1. Sensors
8.7.2. Model choice
8.7.3. Model deployment
8.7.4. Model adaptation
8.8. Future work
8.9. References
List of Authors
Index
A
B, C
D, F
G
H, I
K, L
M
N, O
P, R
S, T
V
WILEY END USER LICENSE AGREEMENT
