Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Grey Wolf Optimizer

The Whale Optimization Algorithm

This paper proposes a novel population-based optimization algorithm called Sine Cosine Algorithm (SCA) for solving optimization problems. The SCA creates multiple initial random candidate solutions and requires them to fluctuate outwards or towards the best solution using a mathematical model based on sine and cosine functions. Several random and adaptive variables also are integrated to this algorithm to emphasize exploration and exploitation of the search space in different milestones of optimization. The performance of SCA is benchmarked in three test phases. Firstly, a set of well-known test cases including unimodal, multi-modal, and composite functions are employed to test exploration, exploitation, local optima avoidance, and convergence of SCA. Secondly, several performance metrics (search history, trajectory, average fitness of solutions, and the best solution during optimization) are used to qualitatively observe and confirm the performance of SCA on shifted two-dimensional test functions. Finally, the cross-section of an aircraft's wing is optimized by SCA as a real challenging case study to verify and demonstrate the performance of this algorithm in practice. The results of test functions and performance metrics prove that the algorithm proposed is able to explore different regions of a search space, avoid local optima, converge towards the global optimum, and exploit promising regions of a search space during optimization effectively. The SCA algorithm obtains a smooth shape for the airfoil with a very low drag, which demonstrates that this algorithm can highly be effective in solving real problems with constrained and unknown search spaces. Note that the source codes of the SCA algorithm are publicly available at http://www.alimirjalili.com/SCA.html .

SCA: A Sine Cosine Algorithm for solving optimization problems

Salp Swarm Algorithm

The paper studied by Daga and Deo (2008) considers the feasibility of using genetic programming (GP) for estimating wind from waves by inverse modeling. The paper includes some problems about the fundamental aspects and use of the proposed GP approach for the aim of their study. In this discussion, some controversial points of the paper are given.

Discussion on “Alternative data-driven methods to estimate wind from waves by inverse modeling” by Mansi Daga, M. C. Deo [Natural Hazards (2008) NHAZ 524, Article 9299, DOI 10.1007/s11069-008-9299-2]

Although obtaining data regarding existing airport pavement condition is an essential task for pavement maintenance, pavement condition surveys can be costly and labor intensive. The objective of this study was to investigate the use of an Android smartphone application as a potential low cost and efficient tool to assess airport pavement condition. The smartphone application (app) records vehicle cab acceleration data, a timestamp, and GPS coordinates. The approximate pavement profile is then back-estimated from the acceleration data using an inverse state space model. The model considers the physics of the mass-spring-damper system of the vehicle sprung mass. The analyses were performed using a MATLAB script to calculate the IRI values.

Pavement Roughness Measurement Using Android Smartphones: Case Study of Missouri Roads and Airports

Rutting continues to be one of the principal distresses in asphalt pavements worldwide. This type of distress is caused by permanent deformation and shear failure of the asphalt mix under the repetition of heavy loads. The Hamburg wheel tracking test (HWTT) is a widely used testing procedure designed to accelerate, and to simulate the rutting phenomena in the laboratory. Rut depth, as one of the outputs of the HWTT, is dependent on a number of parameters related to mix design and testing conditions. This study introduces a new model for predicting the rutting depth of asphalt mixtures using a deep learning technique - the convolution neural network (CNN). A database containing a comprehensive collection of HWTT results was used to develop a CNN-based machine learning prediction model. The database includes 10,000 rutting depth data points measured across a large variety of asphalt mixtures. The model has been formulated in terms of known influencing mixture variables such as asphalt binder high temperature performance grade, mixture type, aggregate size, aggregate gradation, asphalt content, total asphalt binder recycling content, and testing parameters, including testing temperature and number of wheel passes. A rigorous validation process was used to assess the accuracy of the model to predict total rut depth and the HWTT rutting curve. A sensitivity analysis is presented, which evaluates the effect of the investigated variables on rutting depth predictions by the CNN model. The model can be used as a tool to estimate the rut depth in asphalt mixtures when laboratory testing is not feasible, or for cost saving, pre-design trials.

A Deep Learning Approach to Predict Hamburg Rutting Curve

This paper presents a new surface sensing approach for health monitoring of asphalt concrete (AC) pavements utilizing a new class of self-powered wireless sensors. The proposed method was based on the interpretation of the data stored in the memory gates of the sensor. A three-dimensional finite element analysis was performed to obtain the dynamic strain at the surface of the pavement for different damage scenarios. Damage states were defined using the element weakening method. The sensor output data was generated from the time-history of the surface strains. Thereafter, the sensor data was fitted to a Gaussian mixture model (GMM) in order to define an initial damage indicator features. Finally, probabilistic neural network classification scheme was used to classify the damage states. The results indicate that the proposed method is effective in detecting and classifying bottom-up cracks in AC pavements using a surface-mounted network of sensors. Figure 1. Prototype of the SWS. ce vehicle could be effectuated using a Radio Frequency Identification (RFID) scanner to read the data stored on board the memory cells of the sensor. Previous studies on the self-powered wireless sensor showed that cracks in pavement could be detected based on the interpretation of the data of a constant injection rate class of SWS (Chatti et al., 2016). In their study, the SWS was embedded inside the asphalt layer. However, the device could be damaged and their replacement might be expensive. Therefore, placing the sensors network near the top surface of the pavement seems to be an attractive solution. In addition, for the case of a constant injection rate SWS, the data can be fitted to a cumulative density function (CDF). However, in this paper, each gate of the sensor has a specific injection rate, which makes the interpretation of the data more complicated. This study proposes a new method for pavement health monitoring based on a surface sensing approach. The proposed detection mechanism is based on integrating the finite element method (FEM) and probabilistic neural networks (PNN). Intensive finite element (FE) analysis of a moving load on a pavement section was performed to obtain a realistic response. The proposed method uses features extracted from the sensors output distributions to define initial damage indicators. Thereafter, the extracted features from the WSN were fused to increase the classification accuracy. 2 SMART SENSOR AND PROPOSED DAMAGE DETECTION SYSTEM The smart sensor is capable of continuously monitoring the strain events within the host structure. As mentioned before, the memory cells records the cumulative drop of voltage/strain at a preselected threshold level. A schematic representation of the working principle of the sensor is presented in Figure 2. Figure 2(a) represents the input signal and Figure 2(b) displays the output of the sensor. The recorded strain droppage is a function of the cumulative time intersections and the gates injection rates as follow: Sj = S0 − Igj × ∑ Ti j j=1:7 (1) Where Sj is the sensor strain at gate j, Ti j is the duration of time intersection number i at the preselected threshold j (see Figure 2), and Igj is the injection rate of gate j. The injection rates are property of the sensors and they control the strain/voltage droppage rates over time. As seen in Figure 2, there is a considerable loss of the sensed information because the data is compressed as a function of the cumulative time. Therefore, a statistical method was proposed to extract valuable features from the sensor distribution. In this paper, the output histogram is fitted to a Gaussian mixture model. GMMs are powerful tools to describe many types of data. The probability density function (PDF) of a Gaussian mixture (GM) distribution is given by the following expression: p(x) = ∑ ck √2 π σk 2 M k=1 exp [− 1 2 ( x− μk σk ) 2 ] (2)

Amir H. Alavi

Papers

Discussion on “Alternative data-driven methods to estimate wind from waves by inverse modeling” by Mansi Daga, M. C. Deo [Natural Hazards (2008) NHAZ 524, Article 9299, DOI 10.1007/s11069-008-9299-2]

Pavement Roughness Measurement Using Android Smartphones: Case Study of Missouri Roads and Airports

A Deep Learning Approach to Predict Hamburg Rutting Curve

Continuous health monitoring of asphalt concrete pavements using surface-mounted battery-free wireless sensors

Artificial Intelligence in STEM Education