Jelena Velimirovic

Mrs. Jelena Velimirovic

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 August 2021 in Symmetry
Verified
Journal Article
Open access
PDF file
LVLazar Z. Velimirović
Lazar Z. Velimirović
RBRadmila Janković Babić
Radmila Janković Babić
JVJelena Velimirovic
Jelena Velimirovic
AJAleksandar Janjic
Aleksandar Janjic
One way to optimize wastewater treatment system infrastructure, its operations, monitoring, maintenance and management is through development of smart forecasting, monitoring and failure prediction systems using machine learning modeling. The aim of this paper was to develop a model that was able to predict a water pump failure based on the asymmetrical type of data obtained from sensors such as water levels, capacity, current and flow values. Several machine learning classification algorithms were used for predicting water pump failure. Using the classification algorithms, it was possible to make predictions of future values with a simple input of current values, as well as predicting probabilities of each sample belonging to each class. In order to build a prediction model, an asymmetrical type dataset containing the aforementioned variables was used.

21 April 2021 in Symmetry
Verified
Journal Article
Open access
PDF file
JVJelena Velimirovic
Jelena Velimirovic
AJAleksandar Janjic
Aleksandar Janjic
This paper deals with uncertainty, asymmetric information, and risk modelling in a complex power system. The uncertainty is managed by using probability and decision theory methods. More specifically, influence diagrams—as extended Bayesian network functions with interval probabilities represented through credal sets—were chosen for the predictive modelling scenario of replacing the most critical circuit breakers in optimal time. Namely, based on the available data on circuit breakers and other variables that affect the considered model of a complex power system, a group of experts was able to assess the situation using interval probabilities instead of crisp probabilities. Furthermore, the paper examines how the confidence interval width affects decision-making in this context and eliminates the information asymmetry of different experts. Based on the obtained results for each considered interval width separately on the action to be taken over the considered model in order to minimize the risk of the power system failure, it can be concluded that the proposed approach clearly indicates the advantages of using interval probability when making decisions in systems such as the one considered in this paper.

1 January 2023 in Facta universitatis - series: Electronics and Energetics
Verified
Journal Article
Open access
PDF file
AJAleksandar Janjic
Aleksandar Janjic
LVLazar Z. Velimirović
Lazar Z. Velimirović
JVJelena Velimirovic
Jelena Velimirovic
Home energy management systems (HEMS) are increasingly used as a tool that creates optimal consumption and production schedules for Smart Grids, by considering objectives such as energy costs, environmental concerns, load profiles, and consumer comfort. Multiple criteria selection of optimal HEMS seems to be superior to the traditional cost benefit assessment in measuring intangibles and soft impacts, introducing qualitative aspects in the analysis. This paper proposes an algorithm for the selection of optimal HEMS, using the fuzzy AHP method. This methodological framework provides a multi-criteria approach for estimating the benefits and costs of different HEMS within the Smart Grid uncertain environment. This method allows the decision makers to incorporate unquantifiable, asymmetrical, incomplete, nonobtainable information and partially ignorant facts into a decision model. Four criteria and eleven performances for the optimal solution selection are defined. The method is successful in the evaluation of alternatives in the presence of heterogeneous criteria and uncertain environment. The methodology is illustrated on the choice of HEMS from the power distribution company perspective. It is concluded that the evaluation of weighting factors has a decisive character in the choice of the final one of several alternative variants. Fuzzification of input values can also contribute to a more flexible view of the given problem and analysis of sensitivity to various input parameters.

25 November 2025
Verified
Conference Paper
Closed access
JVJelena Velimirovic
Jelena Velimirovic
LVLazar Z. Velimirović
Lazar Z. Velimirović
IPIvana Petkovski
Ivana Petkovski
This paper proposes a data-driven approach to predict session-level EV energy demand using contextual station data and user behavior from historical logs. A large realistic dataset and a time-aware, user-level anti-leakage split ensured robust evaluation. The XGBoost model outperformed other regressors, offering accurate forecasts without vehicle telemetry. This scalable method enables operators to predict energy demand reliably, supporting smarter EV charging network management and allowing future integration of detailed user or technical data.

12 November 2026 in AI
Verified
Journal Article
Open access
PDF file
ABAleksa Bogdanović
Aleksa Bogdanović
LVLazar Z. Velimirović
Lazar Z. Velimirović
JVJelena Velimirovic
Jelena Velimirovic
Zoran StajićZoran Stajić
Zoran Stajić
Accurate prediction of multivariate datasets increasingly relies on deep learning models, often increasing complexity and computing requirements, particularly for multi-horizon predictions. This work proposes a model designed to significantly reduce computing requirements while retaining predictive accuracy. By replacing standard convolutions with depthwise and grouped pointwise operations, the model preserves temporal processing and captures dependencies without substantially complicating the architecture. Evaluated on five multivariate datasets across four forecasting horizons, the proposed model is compared against conventional and depthwise Temporal Convolutional Networks; lightweight variants, such as TSMixer and FITS; and advanced models like Informer, Autoformer, and iTransformer. It reduces multiply–accumulate operations by 58.9–70.8% and trainable parameters by 14.15–35.8% compared to standard temporal convolution networks. Notably, forecasting accuracy varies across datasets and horizons, with small differences between TCN variants corresponding to run-to-run variability. The findings reveal that the model excels in accuracy–efficiency trade-offs, offering an alternative for multivariate time-series forecasting when computational cost is a critical consideration.