Forest fires are a persistent environmental issue in West Kalimantan, Indonesia, driven by both natural and human factors. Fire Radiative Power (FRP) serves as a vital indicator for assessing wildfire intensity and energy release. This study aims to model and predict the spatial temporal dynamics of FRP using the Generalized Space Time Autoregressive [GSTAR(1;1)] model combined with Ordinary Kriging interpolation. The dataset covers West Kalimantan from July 2024 to September 2025, comprising four attributes: observation date, longitude, latitude, and FRP value. Data filtering was applied from the national to provincial level, focusing on three regencies Sanggau, Sekadau, and Ketapang across 14 sub-districts represented by a 1.25×1.25 grid. The data consisted of 65 weekly observations, with 61 used for training and 4 for testing. The GSTAR(1;1) model with a spatial area-based framework achieved an optimal MAPE of 12.63% and satisfied the white noise assumption, indicating reliable performance. Predictions for October 2025 indicated relatively stable fire intensity, with a slight FRP decrease in Nanga Tayap and Sandai during the final week. Overall, the integrated GSTAR–Kriging framework effectively captured both temporal and spatial variations, supporting improved fire risk assessment and regional decision making for wildfire management in West Kalimantan.

The stunting rate in West Kalimantan has reached 27%, mainly due to the government's inability to prioritise regions for essential services and education, especially for adolescents and pregnant women. This study aims to explain the role of modified K-Means and CHI methods in forming optimal clusters and interpreting their conditions. Eight research variables, sourced from BPS and SIGA in 2023, were used: number of adolescents receiving counselling, informed consents, complication cases, aslokon expenditure, aslokon stock, population growth rate, population density, and life expectancy. These variables were validated using PCA with the Kaiser and PVE approaches. Clustering was done by analysing the data for each variable and the characteristics of the objects using the Euclidean distance, determining the centroid values, and iterating until the results stabilised. The clusters were evaluated from one to seven to find the optimal amount using CHI. The results identified five clusters: cluster 1 (relatively poor, three objects), cluster 2 (inferior, four objects), cluster 3 (good, three objects), cluster 4 (exquisite, three objects) and cluster 5 (good, one object).

Multivariate time series forecasting plays a crucial role in various domains, including finance, where accurate stock price prediction supports strategic decision-making. Traditional methods such as Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing (ETS), and Vector Autoregression (VAR) often fall short when dealing with complex, non-linear data—particularly those exhibiting long-term temporal dependencies. This study evaluates deep learning approaches, namely Multilayer Perceptron (MLP), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM), using daily AAPL stock price data from January 2020 to November 2024. The results show that the MLP model with a 10-day time window achieves the best accuracy, yielding lower values in Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE) compared to CNN, LSTM, and VAR. The findings suggest that MLP is particularly effective in capturing complex patterns in multivariate time series forecasting.

The transportation of goods and services is a strategic issue in logistics systems, particularly in the palm oil industry. One of the key distribution optimization challenges is the Capacitated Vehicle Routing Problem (CVRP), which involves determining optimal distribution routes while considering vehicle capacity constraints. This study aims to identify the shortest distribution routes for transporting fresh oil palm fruit bunches from collection points to the palm oil mill, with the goal of minimizing total vehicle travel distance. A heuristic approach using the Saving Matrix method and a metaheuristic approach using a Genetic Algorithm were applied separately to two regions: Block P and Block Q, each consisting of 14 collection points with daily distribution schedules. The performance of both algorithms was analyzed and compared in the context of region-based distribution.The results show that the Genetic Algorithm yields more optimal solutions than the Saving Matrix, reducing the total travel distance by 33.92% in Block P and 32.81% in Block Q. In comparison, the Saving Matrix achieved reductions of 38.72% in Block P and 35.25% in Block Q. These findings indicate that the Genetic Algorithm performs better in solving CVRP for the distribution of fresh oil palm fruit bunches and can serve as a foundation for developing more efficient distribution systems using heuristic and metaheuristic approaches

This study aims to identify and analyze the application of the Mean-Conditional Value-at-Risk (Mean-CVaR) model in the allocation of financial asset portfolio weights combined with the K-Means Clustering algorithm. The Systematic Literature Review (SLR) method is used with the PRISMA protocol through the stages of identification, screening, eligibility, and inclusion. Data is obtained from Scopus, ScienceDirect, and Dimensions databases, then selected up to six relevant primary articles. The results of the study indicate that CVaR is the dominant risk measure in portfolio optimization, while K-Means Clustering serves as a method of grouping assets to increase diversification. The optimization methods used include Genetic Algorithm, Particle Swarm Optimization, Teaching Learning-Based Optimization, and Stochastic Programming. However, direct integration between Mean-CVaR and K-Means within a portfolio weight allocation framework is still rare. This research emphasizes the need to develop a hybrid model that combines both approaches in an integrated manner, applied to a multi-asset portfolio, and validated under various market conditions to produce an optimal, adaptive, and resilient investment strategy against extreme risks.

Although the Durand-Kerner method is widely used across various fields of computer science, especially in numerical computing, it continues to encounter challenges in locating roots of high-degree polynomials, such as issues with accuracies of roots of the polynomial zeros. Our initial tests and observations on several methods for finding polynomial roots revealed that the roots' accuracy starts to degrade noticeably for polynomials where the degree exceeds 10. Based on considerations of algebraic concepts involving polynomial vector spaces, we introduce an improvement of the Durand-Kerner algorithm aimed at improving root precision. This approach includes targeted refinements in coefficient evaluation, identification of root types, and iterative polishing techniques. We also conducted a comparative evaluation to assess its effectiveness against the original Durand Kerner method and MATLAB's roots() function. Overall, the enhanced algorithm delivers superior accuracy for complex roots—particularly in cases involving multiple zero or integer roots—outperforming both benchmarks, but its execution time increases substantially with polynomial degree.

This study investigates the dynamical behavior of a three species ecological system involving unilateral interactions of commensalism and amensalism with Beddington–DeAngelis functional responses. The positivity, boundedness, existence, and uniqueness of the model solutions are established, and four equilibrium points are identified. Stability analysis shows that the coexistence equilibrium point and the neutral species only equilibrium point are locally asymptotically stable, whereas the other equilibria are always unstable. Numerical simulations are conducted to confirm the analytical findings. Ecologically, the results indicate that stability can be achieved only when all neutral species coexist, even though without the commensal–amensal species. In contrast, the commensal-amensal species cannot persist without the presence of all neutral species.

This study develops an integrated framework for pricing double barrier options under time-varying interest rates by combining ARIMA-based forecasting with Monte Carlo simulations. Monthly U.S. Treasury Bill rates from 2019–2025 are modeled using the ARIMA(2,2,0) process to generate dynamic risk-free rates, which are incorporated into three Monte Carlo approaches standard, antithetic variate, and control variate. Tesla Inc. stock prices are used as the underlying asset modeled through Geometric Brownian Motion. The integration of ARIMA-based dynamic rates within the Monte Carlo framework enables more realistic pathwise discounting and improves simulation convergence. The results show that the control variate method provides the most accurate and stable estimates for knock-in call options, whereas the antithetic variate technique yields superior accuracy for knock-in put, knock-out call, and knock-out put options. Overall, the combined use of ARIMA-forecasted interest rates and variance-reduction techniques enhances the precision and stability of double barrier option valuation under dynamic financial conditions.

In food industries, supplier evaluation and selection are strategic activities that influence product freshness, operational continuity, and supply chain sustainability. However, this process is often hindered by uncertainty and ambiguity in expert judgments. In response to these challenges, the present study proposes an integrated decision-making method that combines Circular Intuitionistic Fuzzy Set (CIFS), the Stepwise Weight Assessment Ratio Analysis (SWARA), and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). CIFS capture uncertainty in expert opinions, SWARA determines systematic criteria weights, and TOPSIS—enhanced with the Garg et al. distance measure—ranks suppliers based on aggregated evaluations. The evaluation involves seven key criteria: flexibility, capacity, quality, service, reputation, price, and lead time, assessed across five potential suppliers. Applied to Toko Pia Cap Mangkok, a traditional snack producer in Malang, Indonesia, the method identifies lead time, capacity, and reputation as the most critical criteria. Among the alternatives, Supplier $A_1$ consistently ranks first across optimistic, pessimistic, and combined scenarios, confirming its robustness and reliability, followed by Supplier $A_2$, while others perform less competitively. This study advances fuzzy-based multi-criteria decision-making by integrating CIFS–SWARA–TOPSIS, ensuring reliable supplier selection under uncertainty and offering a replicable framework for decision-makers in the food industry.

Waste generation exceeding landfill capacity highlights the urgency of realizing its economic value. This study analyzes the effect of Quality of Facilities and Infrastructure (X1) and Use of Waste Banks (X2) on Waste Management-Based 3R (Y1) and Waste Economic Value Utilization (Y2) using a truncated spline nonparametric path model. This study evaluates the performance of a nonparametric path analysis model based on truncated spline combined with a double resampling. Data were collected using a Likert scale questionnaire on community perceptions of waste’s economic benefits in Batu City. Simulation results show that the Jackknife-Bootstrap method achieves the lowest average bias (0.058), outperforming single resampling approaches such as Single-Bootstrap (0.178) and Single-Jackknife (0.176). Empirical findings indicate that improvements in the Quality of Facilities and Infrastructure (X1) and Waste Bank Use (X2) significantly enhance Waste Management Based 3R (Y1) and Utilization of Waste Economic Value (Y2). The truncated spline model reveals a saturation effect, where the marginal benefits of X1 and X2 decrease beyond a threshold. Furthermore, Y1 positively affects Y2, emphasizing the importance of efficient waste management in enhancing economic value. The results support policies promoting balanced infrastructure development, community empowerment, and institutional innovation for sustainable circular economy implementation.