Application of MGGP, ANN, MHBMO, GRG, and Linear Regression for Developing Daily Sediment Rating Curves

A data-driven relationship between sediment and discharge of a river is among the most erratic relationships in river engineering due to the existence of an inevitable scatter in sediment rating curves. Recently, Multigene Genetic Programming (MGGP), as a machine learning (ML) method, has been proposed to develop data-driven models for various phenomena in the field of hydrology and water resource engineering. The present study explores the capability of MGGP-based models to develop daily sediment ratings of two gauging sites with 30-year sediment-discharge data, which was utilized previously in the literature. The results obtained by MGGP were compared with those achieved by an empirical model and Artificial Neural Network (ANN). The coefficients of the empirical model were calibrated using linear and nonlinear regression models (Generalized Reduced Gradient (GRG) and the Modified Honey Bee Mating Optimization (MHBMO) algorithm). According to the comparative analysis, the mean absolute error (MAE) at the two gauging stations reduced from 516.54 to 519.23 obtained by nonlinear regression to 447.26 and 504.23 achieved by MGGP, respectively. Similarly, all other performance indices indicated the suitability and accuracy of MGGP in developing sediment ratings. Therefore, it was demonstrated that ML-based models, particularly MGGP-based models, outperformed the empirical models for estimating sediment loads.

Control of Magnetic Manipulator Using Reinforcement Learning Based on Incrementally Adapted Local Linear Models

Reinforcement learning (RL) agents can learn to control a nonlinear system without using a model of the system. However, having a model brings benefits, mainly in terms of a reduced number of unsuccessful trials before achieving acceptable control performance. Several modelling approaches have been used in the RL domain, such as neural networks, local linear regression, or Gaussian processes. In this article, we focus on techniques that have not been used much so far: symbolic regression (SR), based on genetic programming and local modelling. Using measured data, symbolic regression yields a nonlinear, continuous-time analytic model. We benchmark two state-of-the-art methods, SNGP (single-node genetic programming) and MGGP (multigene genetic programming), against a standard incremental local regression method called RFWR (receptive field weighted regression). We have introduced modifications to the RFWR algorithm to better suit the low-dimensional continuous-time systems we are mostly dealing with. The benchmark is a nonlinear, dynamic magnetic manipulation system. The results show that using the RL framework and a suitable approximation method, it is possible to design a stable controller of such a complex system without the necessity of any haphazard learning. While all of the approximation methods were successful, MGGP achieved the best results at the cost of higher computational complexity. Index Terms–AI-based methods, local linear regression, nonlinear systems, magnetic manipulation, model learning for control, optimal control, reinforcement learning, symbolic regression.

Simulations of the Concentration Fields of Rosette-Type Multiport Buoyant Discharges Using Combined CFD and Multigene Genetic Programming Techniques

Rosette-type diffusers are becoming popular nowadays for discharging wastewater effluents. Effluents are known as buoyant jets if they have a lower density than the receiving water, and they are often used for municipal and desalination purposes. These buoyant effluents discharged from rosette-type diffusers are known as rosette-type multiport buoyant discharges. Investigating the mixing properties of these effluents is important for environmental impact assessment and optimal design of the diffusers. Due to the complex mixing and interacting processes, most of the traditional simple methods for studying free single jets become invalid for rosette-type multiport buoyant discharges. Three-dimensional computational fluid dynamics (3D CFD) techniques can satisfactorily model the concentration fields of rosette-type multiport buoyant discharges, but these techniques are typically computationally expensive. In this study, a new technique of simulating rosette-type multiport buoyant discharges using combined 3D CFD and multigene genetic programming (MGGP) techniques is developed. Modeling the concentration fields of rosette-type multiport buoyant discharges using the proposed approach has rarely been reported previously. A validated numerical model is used to carry out extensive simulations, and the generated dataset is used to train and test MGGP-based models. The study demonstrates that the proposed method can provide reasonable predictions and can significantly improve the prediction efficiency.

Field-Based Prediction Models for Stop Penalty in Traffic Signal Timing Optimization

Transportation agencies optimize signals to improve safety, mobility, and the environment. One commonly used objective function to optimize signals is the Performance Index (PI), a linear combination of delays and stops that can be balanced to minimize fuel consumption (FC). The critical component of the PI is the stop penalty “K,” which expresses an FC stop equivalency estimated in seconds of pure delay. This study applies vehicular trajectory and FC data collected in the field, for a large fleet of modern vehicles, to compute the K-factor. The tested vehicles were classified into seven homogenous groups by using the k-prototype algorithm. Furthermore, multigene genetic programming (MGGP) is utilized to develop prediction models for the K-factor. The proposed K-factor models are expressed as functions of various parameters that impact its value, including vehicle type, cruising speed, road gradient, driving behavior, idling FC, and the deceleration duration. A parametric analysis is carried out to check the developed models’ quality in capturing the individual impact of the included parameters on the K-factor. The developed models showed an excellent performance in estimating the K-factor under multiple conditions. Future research shall evaluate the findings by using field-based K-values in optimizing signals to reduce FC.

An evolutionary computational method to formulate the response of unbonded concrete overlays to temperature loading

PurposeUnbonded concrete overlays (UBOLs) are commonly used in pavement rehabilitation. The current models included in the Mechanistic-Empirical Pavement Design Guide cannot properly predict the structural response of UBOLs. In this paper, a multigene genetic programming (MGGP) approach is proposed to derive new prediction models for the UBOLs response to temperature loading.Design/methodology/approachMGGP is a promising variant of evolutionary computation capable of developing highly nonlinear explicit models for characterizing complex engineering problems. The proposed UBOL response models are formulated in terms of several influencing parameters including joint spacing, radius of relative stiffness, temperature gradient and adjusted load/pavement weight ratio. Furthermore, linear regression models are developed to benchmark the MGGP models.FindingsThe derived design equations accurately characterize the UBOLs response under temperature loading and remarkably outperform the regression models. The conducted parametric analysis implies the efficiency of the MGGP-based model in capturing the underlying physical behavior of the UBOLs response to temperature loading. Based on the results, the proposed models can be reliably deployed for design purposes.Originality/valueA challenge in the design of UBOLs is that their interlayer effects have not been directly considered in previous design procedures. To achieve better performance predictions, it is necessary to capture the effect of the interlayer in the design process. This study addresses this important issue via developing new models that can efficiently account for the effects of interlayer on the stress and deflections. In addition, it provides an insight into the effect of several parameters influencing the deflections of the UBOLs. From a computing perspective, a powerful evolutionary computation technique is introduced that overcomes the shortcomings of existing machine learning methods.

Assessment of Artificial Intelligence Models for Developing Single-Value and Loop Rating Curves

Estimation of discharge flowing through rivers is an important aspect of water resource planning and management. The most common way to address this concern is to develop stage-discharge relationships at various river sections. Various computational techniques have been applied to develop discharge ratings and improve the accuracy of estimated discharges. In this regard, the present study explores the application of the novel hybrid multigene genetic programming-generalized reduced gradient (MGGP-GRG) technique for estimating river discharges for steady as well as unsteady flows. It also compares the MGGP-GRG performance with those of the commonly used optimization techniques. As a result, the rating curves of eight different rivers were developed using the conventional method, evolutionary algorithm (EA), the modified honey bee mating optimization (MHBMO) algorithm, artificial neural network (ANN), MGGP, and the hybrid MGGP-GRG technique. The comparison was conducted on the basis of several widely used performance evaluation criteria. It was observed that no model outperformed others for all datasets and metrics considered, which demonstrates that the best method may be different from one case to another one. Nevertheless, the ranking analysis indicates that the hybrid MGGP-GRG model overall performs the best in developing stage-discharge relationships for both single-value and loop rating curves. For instance, the hybrid MGGP-GRG technique improved sum of square of errors obtained by the conventional method between 4.5% and 99% for six out of eight datasets. Furthermore, EA, the MHBMO algorithm, and artificial intelligence (AI) models (ANN and MGGP) performed satisfactorily in some of the cases, while the idea of combining MGGP with GRG reveals that this hybrid method improved the performance of MGGP in this specific application. Unlike the black box nature of ANN, MGGP offers explicit equations for stream rating curves, which may be counted as one of the advantages of this AI model.