Model Constraints Research Articles

The public debt sustainability and its determinants in Sub-Sahara African countries

PurposeThis paper aims to answer the following important questions: Is public debt in Sub-Saharan African (SSA) countries sustainable? What are the determinants of public debt sustainability in these countries, and do these determinants exhibit heterogeneity due to regional, natural resource, and income differences among SSA countries?Design/methodology/approachThis study analyzes the public debt sustainability in SSA countries using the theoretical model known as the Present Value Budget Constraint (PVBC) model developed by Hamilton and Flavin (1986), and adopts the econometric testing method proposed by Trehan and Walsh (1991). Moreover, to empirically investigate the determinants of public debt sustainability in SSA countries, the System-Generalized Method of Moments (System-GMM) method is applied. Furthermore, this study conducts heterogeneity analysis by categorizing the sample based on different regions, natural resource endowments, and income levels. The data of this study are sourced from the IMF and World Bank databases for 45 SSA countries from 2005 to 2021.FindingsFindings reveal that public debt in SSA countries is not sustainable in the long run, with factors such as the previous government debt, long-term debt ratio, debt repayment capacity, economic growth rate, inflation rate, export to GDP, and government fiscal deficit rate influencing sustainability. Additionally, the factors exhibit heterogeneity attributed to regional, natural resource, and income variations among SSA countries.Practical implicationsThe findings of our study will serve as a catalyst for policymakers in the SSA countries to embrace and sustain robust fiscal consolidation and debt stabilization measures. Moreover, countries with distinct characteristics should implement tailored approaches. Additionally, policymakers in SSA countries should implement economic measures to address public debt issues. These measures include improving the macroeconomic structure, promoting economic transformation and diversification of industries, fostering sustainable economic growth, ensuring price stability, and strengthening resilience against external shocks and debt risks. Specifically, countries endowed with indigenous species, resources, and tourism potential should adopt a well-coordinated strategy that utilizes agriculture, tourism, ecotourism, and the hospitality industry as instruments for sustainable local community and rural development.Originality/valueFirstly, it assesses the sustainability of public debt and its determinants for countries in SSA, which distinguishes it from previous studies that only focus on either debt sustainability or determinants of debt separately. Secondly, by including multiple SSA countries in the analysis, this study stands out from prior research that predominantly concentrates on specific nations. Thirdly, the utilization of the System-GMM method for analyzing determinants adds a novel dimension to this study, departing from earlier literature primarily focused on debt thresholds. Lastly, the heterogeneity analysis conducted in this study provides an empirical foundation for tailoring policies to different countries, addressing a facet often overlooked in earlier literature.

Robust optimal power flow considering uncertainty in wind power probability distribution

This paper proposes an optimal power flow model that takes into account the uncertainty in the probability distribution of wind power. The model can schedule controllable generators under any possible distribution of wind power to ensure the safe and economic operation of the system. Firstly, considering the incompleteness of historical wind power data, the paper models the uncertainty of wind power using second-order moments of probability distribution and their fluctuation intervals. Subsequently, a robust optimal power flow model based on probability distribution model and joint chance constraints is established. The Lagrangian duality theorem is then employed to eliminate random variables from the optimization model, transforming the uncertainty model into a deterministic linear matrix inequality problem. Finally, a convex optimization algorithm is used to solve the deterministic problem, and the results are compared with traditional chance-constrained optimal power flow model. The feasibility and effectiveness of the proposed method are validated through case study simulations.

Group sparsity residual constraint model with weighted log-sum penalty for image restoration

In the field of image restoration, non-convex penalties and non-local self-similarity (NSS) prior have been found to enhance image edges and textures. However, NSS-based methods often overlook the inherent variations within patch groups by treating them as a single entity for processing. Despite the existence of numerous convex or non-convex regularizers, few of them have been able to adaptively adjust penalty strength based on NSS information in order to preserve fine details in an image. To address this issue, we propose a novel group sparsity residual constraint model with weighted log-sum penalty (GSRC-Log) for image restoration in this paper. The proposed regularization term is parameterized by an adaptive scalar parameter that relates to NSS, allowing the model to prioritize high NSS group sparse coefficients while penalizing low NSS coefficients. For image denoising, we provide detailed analysis on the optimal condition of our proposed model and prove that it has a simple closed-form global solution. For image restoration tasks, we utilize the ADMM algorithm to solve the GSRC-Log problem and empirically verify its convergence. Extensive experimental results on both image denoising and compressive sensing (CS) recovery demonstrate that our proposed GSRC-Log outperforms many popular or state-of-the-art methods.

Probing the dynamical and kinematical structures of detached shells around AGB stars

The chemical evolution of asymptotic giant branch (AGB) stars is driven by repeated thermal pulses (TPs). The duration of a TP is only a few hundred years, whereas an inter-pulse period lasts $10^4-10^5$ years. Direct observations of TPs are hence unlikely. However, the detached shells seen in CO line emission that are formed as a result of a TP provide indirect constraints on the changes experienced by the star during the pulse. We aim to resolve the spatial and kinematic sub-structures in five detached-shell sources to provide detailed constraints for hydrodynamic models that describe the formation and evolution of the shells. We used observations of the emission towards five carbon-AGB stars with ALMA (Atacama Large Millimeter/submillimeter Array), including previously published observations of the carbon AGB star U Ant. The data have angular resolutions of 0 to 1 and a velocity resolution of 0.3\ This enabled us to quantify spatial and kinematic structures in the shells. Combining the ALMA data with single-dish observations of the to emission towards the sources, we used radiative transfer models to compare the observed structures with previous estimates of the shell masses and temperatures. The observed emission is separated into two distinct components: a more coherent, bright outer shell and a more filamentary, fainter inner shell. The kinematic information shows that the inner sub-shells move at a higher velocity relative to the outer sub-shells. The observed sub-structures reveal a negative velocity gradient outwards across the detached shells, confirming the predictions from hydrodynamical models. However, the models do not predict a double-shell structure, and the CO emission likely only traces the inner and outer edges of the shell, implying a lack of CO in the middle layers of the detached shell. Previous estimates of the masses and temperatures are consistent with originating mainly from the brighter subshell, but the total shell masses are likely lower limits. Also, additional structures in the form of partial shells outside the detached shell around V644 Sco, arcs within the shell of R Scl, and a partially filled shell for DR Ser indicate a more complicated evolution of the shells and mass-loss process throughout the TP cycle than previously assumed. The observed spatial and kinematical splittings of the shells appear consistent with results from the hydrodynamical models, provided the CO emission does not trace the density distribution in the shell but rather traces the edges of the shells. The hydrodynamical models predict very different density profiles depending on the evolution of the shells and the different physical processes involved in the wind-wind interaction (e.g. heating and cooling processes). It is therefore not possible to constrain the total shell mass based on the CO observations alone. Additional features outside and inside the shells complicate the interpretation of the data. Complementary observations of, for example, CI as a dissociation product of CO would be necessary to understand the distribution of CO compared to in addition to new detailed hydrodynamical models of the pre-pulse, pulse, and post-pulse wind. Only a comprehensive combination of observations and models will allow us to constrain the evolution of the shells and the changes in the star during the thermal-pulse cycle.

GNET Derived Mass Balance and Glacial Isostatic Adjustment Constraints for Greenland

AbstractMonitoring the Greenland mass balance (GMB) is crucial in the context of global sea level rise. Currently, three main methods are used to measure GMB, with the primary source of uncertainty arising from the glacial isostatic adjustment (GIA) contribution. Here, we propose a novel approach based on a simple methodology that uses the entire Greenland GNSS network (GNET) as an instrument to monitor the present‐day mass changes. Our method is validated against GRACE‐derived GMB, and we find a very good agreement. This leads to an independent methodology for estimating present‐day mass changes from GNSS, bridging the gap between GRACE and GRACE‐FO in GMB estimates. Through a combined analysis of GMB from GRACE and GNET, we identify a consistency relation between the gravity and uplift signature of GIA, providing a new robust constraint for GIA models.

Observational constraints on the stellar recycled gas in active galactic nuclei feeding

Abstract Near-infrared long-slit spectroscopy has been used to study the stellar population (SP) of the low luminosity active galactic nuclei (AGN) and matched analogues (LLAMA) sample. To perform the SP fits we have employed the X-shooter simple stellar population models together with the starlight code. Our main conclusions are: The star formation history of the AGNs is very complex, presenting many episodes of star formation during their lifetimes. In general, AGN hosts have higher fractions of intermediate-age SP (light-weighted mean ages, <t > L ≲ 4.5 Gyr) when compared with their analogues (<t > L ≲ 8.0 Gyr). AGN are more affected by reddening and require significant fractions of featureless continuum and hot dust components. The ratio between the AGN radiated energy and the gravitational potential energy of the molecular gas (ERad/EPG) for the AGN is compared with the <t > L and a possible anti-correlation is observed. This suggests that the AGN is affecting the star formation in these galaxies, in the sense that more energetic AGN (log(ERad/EPG) ≳ 3) tend to host nuclear younger SP (<t > L ≲4Gyr). We found that the recent (t <2 Gyr) returned (recycled) stellar mass is higher in AGN than in the controls. We also provide evidence that the mass loss of stars would be enough to feed the AGN, thus providing observational constraints for models that predict that AGN feeding is partially due to the recycled gas from dying stars.

Numerical algorithm for estimating a conditioned symmetric positive definite matrix under constraints

SummaryWe present RCO (regularized Cholesky optimization): a numerical algorithm for finding a symmetric positive definite (PD) matrix with a bounded condition number that minimizes an objective function. This task arises when estimating a covariance matrix from noisy data or due to model constraints, which can cause spurious small negative eigenvalues. A special case is the problem of finding the nearest well‐conditioned PD matrix to a given matrix. RCO explicitly optimizes the entries of the Cholesky factor. This requires solving a regularized non‐linear, non‐convex optimization problem, for which we apply Newton‐CG and exploit the Hessian's sparsity. The regularization parameter is determined via numerical continuation with an accuracy‐conditioning trade‐off criterion. We apply RCO to our motivating educational measurement application of estimating the covariance matrix of an empirical best linear prediction (EBLP) of school growth scores. We present numerical results for two empirical datasets, state and urban. RCO outperforms general‐purpose near‐PD algorithms, obtaining ‐smaller matrix reconstruction bias and smaller EBLP estimator mean‐squared error. It is in fact the only algorithm that solves the right minimization problem, which strikes a balance between the objective function and the condition number. RCO can be similarly applied to the stable estimation of other posterior means. For the task of finding the nearest PD matrix, RCO yields similar condition numbers to near‐PD methods, but provides a better overall near‐null space.

Composite continuum robots: Accurate modeling and model reduction

Models of continuum robots often trade off accuracy and speed. The constant curvature assumption model can perform fast calculations but with poor accuracy. Static models can obtain good calculation accuracy but with slower computation speed. In this paper, a precise and efficient method for establishing a non-constant curvature kinematic model for composite continuum robots is proposed. First, an algorithm based on the Beam Constraint Model is proposed to establish the static model of the continuum robot. The situation in which friction changes gradually in continuum robots is considered in the static model. Then, the Kepler elliptic curve is used to simplify the static model to build a non-constant curvature kinematic model of the continuum robot. Then, complex inverse kinematics solutions are transformed into explicit sets of equations by mathematical transformation. Finally, simulations and experiments are designed to demonstrate the algorithm proposed in this paper. The simulation results show that the proposed algorithm is 1572 times faster than the Levenberg–Marquardt algorithm. The experimental results show that the average error of the static model is 0.065 mm, accounting for 0.13% of its length. The maximum error is 0.309 mm, accounting for 0.618% of its length. The positioning accuracy of the non-constant curvature model proposed in this paper is 2.02 times that of the constant curvature model. The results show that the simplified kinematic model algorithm based on the static model has high accuracy and computational efficiency. This research is crucial for the practical application of continuum robots in engineering.

Spatial-temporal modeling of oil condition monitoring: A review

Lubricating oil plays a vital role as the information carrier for equipment tribological performance. Therefore, oil condition monitoring (OCM) serves as a crucial technology for assessing and predicting state degradation, providing the first-line defense against functional failure. However, limited by random fluctuation and insufficient knowledge, the modeling of OCM has suffered from uncertainty problems, leading to poor applicability and insufficient generalizability. The existing reviews are less elaborated on the information uncertainty for description, characterization, and treatment, which expresses the essence of modeling constraints. To bridge this gap, the paper provides a comprehensive review of the oil state modeling in terms of uncertainty. The existing methods and metrics are reviewed from the perspective of the spatial oil state assessment (OCA) and temporal remaining useful life (RUL) prediction modeling. Further, the existing methods are analyzed to solve the uncertainty problem, and then the solutions in oil state modeling considering uncertainty are discussed. Finally, targeting the challenges of the monitoring technology, the future trends of OCM are presented.

The association between mixed exposure to per- and polyfluoroalkyl substances (PFAS) and the risk of diabetes

IntroductionThis study aims to investigate the relationship between exposure to per- and polyfluoroalkyl substances (PFAS) and the risk of diabetes by analyzing a large dataset from the National Health and Nutrition Examination Survey (NHANES).Material and methodsThis study analyzed data from the US NHANES database from 2003 to 2018. Generalized linear regression models were used to examine the relationship between serum chemical concentrations and diabetes. Weighted quantiles and regression were employed to assess the association between mixed chemical exposure and diabetes. The 'mixture effect' between chemicals and diabetes was estimated using a Quantile g-Computation model.ResultsThe study involved 11,780 participants, with 1,451 individuals having diabetes and 10,329 being non-diabetic. Logistic regression analysis was to investigate the relationship between five PFAS chemicals and the diabetes risk, while controlling for all covariates. The results indicated that none PFAS chemicals exhibited a statistically significant association with an increased risk of diabetes. However, when exploring the positive and negative constraint models of the Weighted Quantile Sum (WQS), the coefficient of the chemical mixing index displayed a strong correlation with diabetes in model III, even after adjusting for covariates. Additionally, an evaluation of the 'mixture effect' using Quantile g-computation revealed that among the five chemicals, the combined exposure effect of Perfluorononanoate (PFNA) on diabetes was positive, whereas exposure to the other four chemicals had a negative impact.ConclusionsThe study findings suggest a possible connection between PFAS and an increased risk of diabetes, particularly highlighting PFNA's positive correlation with the combined exposure effect on diabetes.

Theoretical research of impact vibration of multi-section serial pipeline inspection gauge in large drop pipeline

When operating in large drop pipelines, pipeline inspection gauge may experience severe impact vibration at elbows, potentially leading to its failure. This paper focuses on serial pipeline inspection gauges with multiple sections in large drop pipelines, analyzes the force acting on the pipeline inspection gauge during the descent phase, and establishes a multi-body dynamics model of the gauge under pipe-soil coupling conditions. In addition, a correction formula is developed for the pressure changes in the medium at the front and rear of the pipeline inspection gauge. To solve the model, MSC/ADAMS and MATLAB/Simulink are used for the bi-directional fluid–structure-interaction joint simulation. Moreover, the results are compared with the pipeline inspection gauge motion model established by OLGA. The study investigates the dynamic responses, vibration superposition and laws of dynamic evolution at the upper and lower elbows of the large drop pipeline. The results show that the speed of the gauge increases suddenly during the descent in large drop sections and then stabilizes, with different effects of the fluid medium pressure difference at various stages. The vibration responses along the axial, horizontal radial, and vertical directions at the elbows vary, with the second section experiencing stronger vibrations due to superposition effects. Compared to fixed boundaries, the acceleration extremes of the gauge are lower under the constraints of pipe-soil model, indicating significant soil buffering and vibration damping effects.

Radial Spectral Evolution of the Elsässer Variable z − in the Slow Solar Wind

The nature of the Elsässer variable z − plays a significant role in the understanding of solar wind turbulence. The radial spectral evolution of z − observed by the Helios spacecraft suggests that z − has a persistent background spectrum in the low-frequency range from 0.3 to 1 au. However, the radial evolution in the near-Sun solar wind remains unknown. Here we use Parker Solar Probe measurements to analyze the averaged spectra of z − in the slow solar wind from 0.1 to 0.5 au and compare the results with the WIND observations at 1 au. We find that the averaged spectra of z − do not show the low-frequency break at the low-frequency breakpoint on the power spectra of z +, magnetic field, and velocity. The energy in the frequency range of 10−4–10−2 Hz demonstrates a radial decay from 0.1 to 0.5 au and keeps nearly constant to 1 au. The spectral index of z − gradually decreases from −1.27 at 0.1–0.2 au to −1.59 at 0.4–0.5 au and then slowly declines to −1.70 at 1 au. The radial behaviors of the ∥, ⊥1, and ⊥2 components of z − in the field-aligned coordinates are similar to that of the trace of z −. Our findings suggest a faster radial spectral evolution of z − in the near-Sun solar wind and provide strong observational constraints for the solar wind turbulence model.

Ultralocal Model-Free Adaptive Supertwisting Nonsingular Terminal Sliding Mode Control for Magnetic Levitation System

The magnetic levitation technology has attracted blooming attention from all walks of life thanks to its superiorities such as non-contact and active control. While the intrinsic features of open-loop instability and strong nonlinearity, along with the uncertain external disturbance make the levitation control particularly significant and difficult. For speeding up the convergence and obtaining a better stability and disturbance robustness, this paper proposes an ultra-local model-free adaptive super-twisting nonsingular terminal sliding mode control (ULMF-ASTNTSMC) strategy for the magnetic levitation system. Different from the conventional model-based control, ultra-local model-free control breaks the constraint of a precise system model. Therefore, an improved second-order ultra-local model is established for the magnetic levitation system. For the benefit of finite time convergence and non-singularity, the nonsingular terminal sliding mode surface is chosen. As a second-order algorithm, super-twisting (ST) algorithm can fundamentally suppress the chattering effect of the sliding mode control (SMC). An adaptive super-twisting algorithm is designed for the further improvement of the system dynamic response. The stability of the proposed scheme is strictly verified based on the Lyapunov stability theory. Simulation and experimental comparisons with the existing control methods validate the exceeding performance of the proposed strategy under various operating conditions.

Energy flux couples sulfur isotope fractionation to proteomic and metabolite profiles in Desulfovibrio vulgaris.

Microbial sulfate reduction is central to the global carbon cycle and the redox evolution of Earth's surface. Tracking the activity of sulfate reducing microorganisms over space and time relies on a nuanced understanding of stable sulfur isotope fractionation in the context of the biochemical machinery of the metabolism. Here, we link the magnitude of stable sulfur isotopic fractionation to proteomic and metabolite profiles under different cellular energetic regimes. When energy availability is limited, cell-specific sulfate respiration rates and net sulfur isotope fractionation inversely covary. Beyond net S isotope fractionation values, we also quantified shifts in protein expression, abundances and isotopic composition of intracellular S metabolites, and lipid structures and lipid/water H isotope fractionation values. These coupled approaches reveal which protein abundances shift directly as a function of energy flux, those that vary minimally, and those that may vary independent of energy flux and likely do not contribute to shifts in S-isotope fractionation. By coupling the bulk S-isotope observations with quantitative proteomics, we provide novel constraints for metabolic isotope models. Together, these results lay the foundation for more predictive metabolic fractionation models, alongside interpretations of environmental sulfur and sulfate reducer lipid-H isotope data.

Dynamically driven correlations in elastic net models reveal sequence of events and causality in proteins.

An explicit analytic solution is given for the Langevin equation applied to the Gaussian Network Model of a protein subjected to both a random and a deterministic periodic force. Synchronous and asynchronous components of time correlation functions are derived and an expression for phase differences in the time correlations of residue pairs is obtained. The synchronous component enables the determination of dynamic communities within the protein structure. The asynchronous component reveals causality, where the time correlation function between residues i and j differs depending on whether i is observed before j or vice versa, resulting in directional information flow. Driver and driven residues in the allosteric process of cyclophilin A and human NAD-dependent isocitrate dehydrogenase are determined by a perturbation-scanning technique. Factors affecting phase differences between fluctuations of residues, such as network topology, connectivity, and residue centrality, are identified. Within the constraints of the isotropic Gaussian Network Model, our results show that asynchronicity increases with viscosity and distance between residues, decreases with increasing connectivity, and decreases with increasing levels of eigenvector centrality.

Insomnia-related rodent models in drug discovery.

Despite the widespread prevalence and important medical impact of insomnia, effective agents with few side effects are lacking in clinics. This is most likely due to relatively poor understanding of the etiology and pathophysiology of insomnia, and the lack of appropriate animal models for screening new compounds. As the main homeostatic, circadian, and neurochemical modulations of sleep remain essentially similar between humans and rodents, rodent models are often used to elucidate the mechanisms of insomnia and to develop novel therapeutic targets. In this article, we focus on several rodent models of insomnia induced by stress, diseases, drugs, disruption of thecircadian clock, and other means such as genetic manipulation of specific neuronal activity, respectively, which could be used to screen for novel hypnotics. Moreover, important advantages and constraints of some animal models are discussed. Finally, this review highlights that the rodent models of insomnia may play a crucial role in novel drug development to optimize themanagement of insomnia.

Machine vision model for detection of foreign substances at the bottom of empty large volume parenteral.

Empty large volume parenteral (LVP) bottle has irregular shape and narrow opening, and its detection accuracy of the foreign substances at the bottom is higher than that of ordinary packaging bottles. The current traditional detection method for the bottom of LVP bottles is to directly use manual visual inspection, which involves high labor intensity and is prone to visual fatigue and quality fluctuations, resulting in limited applicability for the detection of the bottom of LVP bottles. A geometric constraint-based detection model (GCBDM) has been proposed, which combines the imaging model and the shape characteristics of the bottle to construct a constraint model of the imaging parameters, according to the detection accuracy and the field of view. Then, the imaging model is designed and optimized for the detection. Further, the generalized GCBDM has been adopted to different bottle bottom detection scenarios, such as cough syrup and capsule medicine bottles by changing the target parameters of the model. The GCBDM, on the one hand, can avoid the information at the bottom being blocked by the narrow opening in the imaging optical path. On the other hand, by calculating the maximum position deviation between the center of visual inspection and the center of the bottom, it can provide the basis for the accuracy design of the transmission mechanism in the inspection, thus further ensuring the stability of the detection.

Robust elastic full-waveform inversion using an alternating direction method of multipliers with reconstructed wavefields

Elastic full-waveform inversion (EFWI) is a process used to estimate subsurface properties by fitting seismic data while satisfying wave propagation physics. The problem is formulated as a least-squares data fitting minimization problem with two sets of constraints: partial-differential equation (PDE) constraints governing elastic wave propagation and physical model constraints implementing prior information. The alternating direction method of multipliers is used to solve the problem, resulting in an iterative algorithm with well-conditioned subproblems. Although wavefield reconstruction is the most challenging part of the iteration, sparse linear algebra techniques can be used for moderate-sized problems and frequency domain formulations. The Hessian matrix is blocky with diagonal blocks, making model updates fast. Gradient ascent is used to update Lagrange multipliers by summing PDE violations. Various numerical examples are used to investigate algorithmic components, including model parameterizations, physical model constraints, the role of the Hessian matrix in suppressing interparameter cross-talk, computational efficiency with the source sketching method, and the effect of noise and near-surface effects.

Spatial modeling of land resources and constraints to guide urban development in Saudi Arabia’s NEOM region using geomatics techniques

This research presents a GIS-based approach to evaluate land suitability for the urbanization of Saudi Arabia’s NEOM region, an ambitious initiative for sustainable development within a desert setting. By employing a multi-criteria decision-making framework, we have synthesized data on land resources, stability, accessibility, construction costs, and proximity to conservation areas using the Analytical Hierarchy Process (AHP) to prioritize these factors based on their importance for sustainable growth. Our analysis reveals that areas designated with grid values of 9 and 8, totaling 6312.33 square kilometers or 28.23% of the area studied, are most favorable for urban development. These sectors span the northern, central, and southern parts of NEOM, which are notable for their strategic location near future infrastructure and natural assets conducive to sustainable practices. This finding supports NEOM’s key initiatives such as “The Line,” “Oxagon,” and “TROJENA,” showcasing the model’s utility in directing urban development strategies. The study discusses the broader socio-economic and environmental ramifications of pinpointing areas suited and unsuited for development, underlining the model’s role in facilitating decision-making that aligns urban expansion with environmental stewardship. Furthermore, it advocates for leveraging development-restricted zones in conservation efforts and renewable energy ventures. Nevertheless, the study concedes certain limitations, such as the dependence on static datasets and inherent challenges related to the AHP method, including the presumption of criteria independence. Future research avenues are proposed, emphasizing the integration of dynamic data sources, comprehensive socio-cultural impact evaluations, and adopting flexible urban planning methodologies to refine the model’s precision and practicality. Overall, this detailed examination showcases NEOM’s blueprint as an exemplar for future urban environments, illustrating a systematic strategy for developing desert regions that integrate innovation, ecological sustainability, and the preservation of cultural heritage.

Spatial constraints in cellular automata-based urban growth models: A systematic comparison of classifiers and input urban maps

Spatial constraints are fundamental to integrating the spatial suitability to urbanization into Cellular Automata-based (CA) urban growth models, but there is a lack of consensus on the optimal methods for this purpose. This study compared the performance of three probabilistic classifiers to generate suitability surfaces for CA-based urban growth models: Logistic Regression using Generalized Linear Model (LR-GLM), Logistic Regression using Generalized Additive Model (LR-GAM), and Random Forest (RF). The study also evaluated the sensitivity of these classifiers to the input urban map adopted as a dependent variable. For this analysis, seven maps were tested: the historical urban map containing the entire extent of the urban footprint, and six additional maps containing only the recently urbanized areas over timeframes ranging from one year up to two decades. The comparison evaluated the goodness of fit of the suitability surfaces and the spatial accuracy of the urban growth simulations, using five large Brazilian cities as case study areas. The results revealed that the RF classifier significantly outperformed the LR-based classifiers. However, this overperformance was more prominent when incorporating the new urban cells over the last one to two decades of growth as input urban maps. In addition, the sensitivity analysis of the input urban maps emphasized the benefits of calibrating the classifier using the recently urbanized cells rather than the historical urban extent. We consistently observed these results concerning classifiers and input urban maps across all five case study areas. Thus, the RF classifier combined with a training dataset containing the newly urbanized areas over at least the last 10 years systematically resulted in the suitability surfaces with the highest predictability among all tested scenarios.