Application Study of Image Processing Transfer Learning Based on TOPSIS Algorithm
in Water Quality Resource Monitoring and Evaluation Model
(Jingjing Lan)
1
(Peng Zhang)
1,*
(Lingjun Wu)
1
(Shuangshuang Huang)
1
(Jinxi Zhang)
1
-
(School of Hydraulic Engineering, Wanjiang University of Technology, Ma’anshan 243000,
China)
Copyright © The Institute of Electronics and Information Engineers(IEIE)
Keywords
TOPSIS, Water quality testing, Transfer learning, Image processing
1. Introduction
Environmental issues, particularly pollution, are closely tied to people’s livelihoods.
Water, a vital component of the ecosystem, is often considered the source of life.
Despite covering about 70% of the Earth’s surface, only 2.5% of water resources are
fresh and usable, with the majority locked in polar ice caps [1,
2]. In China, for example, per capita fresh water resources declined in 2021, and a
significant portion of its water environments are polluted, primarily due to high
levels of chemical oxygen demand, elevated acid salt index, and total phosphorus.
As China’s economy and society advance, water demand and pollution—especially from
urban sewage and industrial waste—have increased, leading to numerous major water
pollution incidents in recent years [3]. Pollution is mainly driven by human activities, including land-based domestic sewage,
industrial discharge, and land reclamation. These pollutants can directly or indirectly
degrade water bodies, damaging the water environment and impairing its ability to
provide clean drinking water [4,
5]. To ensure the sustainable development of water resources, China has incorporated
water resource protection into its national strategy, making the prevention and control
of water pollution a global priority. China’s water environment management and control
technology started slowly. In 2021, China carried out pilot investigations in various
river basins. The evaluation results showed that among the 701 points in water basins,
excellent points accounted for 40.1%, good points accounted for 40.8%, and poor and
poor points accounted for 19.1%, among which poor and poor conditions increased by
5.1% compared with 14% in 2020 [6]. From this, it can be seen that some regions in China have problems such as deteriorating
water quality, shortage of water resources, and excessive sewage discharge [7].
Natural rivers, being the primary source of domestic water for numerous cities, are
intricately intertwined with the health of residents. As a result, safeguarding the
water resources and environment of these rivers has emerged as a pivotal research
direction [8]. Water pollution incidents are not just conventional occurrences; they also encompass
unconventional pollution events. Because of their different discharge places, time
and ways, pollutants spread rapidly in a short time, causing serious water pollution.
How to integrate the actual situation of water environment and water use function
in a short time, select appropriate indicators to analyze and monitor the water quality,
and improve the accuracy of water quality anomaly detection. At present, the water
environment monitoring technology needs to be improved [9], and the traditional instrument detection method has a long measurement period and
cumbersome process when judging water quality [10]; When comparing and analyzing the original monitoring data based on the national
standard or empirical threshold, the hidden information contained in the water quality
data cannot be effectively mined and its value can be brought into play; There is
a lag in laboratory sampling analysis and inspection, which cannot solve sudden water
environment problems in time [11]. In view of the shortcomings of the above methods, this paper proposes a method combining
anomaly detection algorithm with neural network. By adopting a reasonable prediction
model for data mining of water quality parameters, the hidden information and correlation
among various parameters can be deeply learned. At the same time, efficient and accurate
water quality prediction can more clearly reflect the current state of water pollution
and the future change trend of water quality. Neural network takes time series data
as training samples, deeply excavates the relationship among various parameters of
water quality, and makes the water quality prediction model have high accuracy by
using the nonlinear characteristics of autonomous learning. In real life, water quality
samples have certain fuzzy characteristics, but most of the traditional anomaly detection
methods belong to hard partition. Employing the unique characteristics of the fuzzy
clustering algorithm and its integration can effectively address the shortcomings
and issues encountered in traditional water quality anomaly detection. Furthermore,
neural networks possess robust adaptability and flexibility, enabling the integration
of various other methodologies into this comprehensive model. Therefore, this study
can quickly and accurately detect water quality anomalies, timely and effectively
adjust various measures to protect the water environment, and improve the increasingly
serious trend of water pollution. At the same time, it is of great significance to
prevent emergencies and improve early warning.
2. Correlative Theory of Water Quality Anomaly Detection
2.1. Water Quality Detection Index
In China, water quality monitoring data is generally collected every 4 hours, though
this frequency can be adjusted in emergencies. Indicators of abnormal river water
quality are categorized into three types: physical, chemical, and biological [12,
13]. Water quality assessment typically starts with evaluating physical indices, which
include sensory metrics such as temperature, color, turbidity, and transparency. These
metrics are crucial as they are directly related to the solubility of organic matter
in the aquatic environment and various transformation processes. Water color comprises
surface color and true color. Surface color results from the combined effects of dissolved,
colloidal, and suspended substances, while true color is the color observed after
removing suspended solids from the surface color. Abnormal color is a key indicator
of water pollution. Turbidity and transparency also play important roles in assessing
water quality.
Fig. 1. TOPSIS algorithm in water quality resource monitoring data preprocessing flowchart.
Fig. 1 shows TOPSIS algorithm in water quality resource monitoring data preprocessing flowchart.
The correlation between transparency and turbidity is pivotal in understanding the
overall condition of aquatic environments and their susceptibility to pollution and
other anthropogenic pressures. The more the amount, the lower its visibility. Other
physical indexes include total solids, suspended solids, fixed solids and volatile
solids, electrical conductivity and so on. Solid matter (TS), solid matter in water
refers to the residual matter after evaporation of water samples at a certain temperature
[14]. Solids in water can be classified into two categories based on their solubility:
dissolved solids (DS) and suspended solids (SS). Dissolved solids, referred to as
total filterable residue, represent the cumulative number of various substances that
are capable of dissolving in water. In water analysis, these solids are identified
as the solid matter obtained through evaporation of water at a high temperature ranging
from 103-105°C. Conversely, suspended solids refer to the total non-filterable residues.
Water samples are filtered through a 0.45-micron filter membrane. Solid particles
that cannot pass through this filter membrane are called suspended solids. The TOPSIS
ideal solution calculation formula and the TOPSIS closeness calculation formula are
shown in Eqs. (1) and (2).
Fixed solids and volatile solids are measured by evaporating and drying water samples
at 600°C. Volatile solids are defined as the mass lost due to combustion, while the
residual solid remaining after burning is referred to as fixed solids [15]. Conductivity measures how well water conducts electricity, which indirectly reflects
the solubility of various parameters, including total solids, salinity, and total
ion concentration—key indicators in water quality detection [16]. Chemical indicators of water quality include pH, Chemical Oxygen Demand (COD), Biochemical
Oxygen Demand (BOD), and Total Organic Carbon (TOC). The pH level indicates acidity
or alkalinity, with a neutral value of 7. COD and BOD measure the oxidation of organic
matter, while TOC quantifies the carbon content in both dissolved and suspended matter.
Biological indicators, such as total bacterial count and coliform bacteria, are also
important. Elevated bacterial counts often signal water pollution, although this indicator
alone cannot pinpoint the pollution source and should be used in conjunction with
tests for coliform bacteria. Coliform bacteria, which are produced from lactose decomposition
within 24 hours at 35°C to 37°C, can help assess the sanitary quality of water [17]. Water quality indices are crucial for analyzing the current state of the water environment
and determining whether the water body is polluted.
2.2. Fluctuation Types of Water Quality Indicators
Abnormal changes in water quality are usually defined as significant changes in one
or more water quality indicators, showing certain fluctuation characteristics with
time changes, such as the following categories:
2.2.1 Background data fluctuation
Conventional water quality monitoring data is usually disturbed by external factors
with great fluctuation, such as temperature, environment, season and other factors,
which will cause periodic fluctuation of signals. The normal fluctuation of background
data is not the occurrence of abnormal events [18,
19]. As a typical time, series data, the influence of background fluctuation can be minimized
by comparing the predicted value of water quality prediction model with the measured
value in data analysis. The image preprocessing formula and the CNN convolution layer
calculation formula are shown in Eqs. (3) and (4).
2.2.2 Noise and outliers
Outliers are distinctively dissimilar from the overall data distribution, predominantly
evidenced by their brief duration of a sudden surge or drop in sampling values at
a given time, followed by a swift return to the anticipated range in the subsequent
period. Typically, data points that do not fall into this outlier category are referred
to as “normal data”, whereas outliers are commonly known as “abnormal data”. Outliers
are different from noise data. Usually, the sum of real data and noise is defined
as observation, while outliers belong to observation. When outliers have anomalies,
it is difficult to distinguish whether they are caused by real data collected by sensors
or affected by noise.
2.2.3 Baseline changes
Baseline changes are often caused by process operations, such as valve opening and
closing, equipment maintenance, water quantity changes, etc. These operations will
cause sudden changes in water quality indicators [20]. Baseline is similar to slowly changing trend, signal drift, instrument drift or
background drift, and the above fluctuation information interference cannot be eliminated
by high-pass filter. If baseline change is not eliminated, the frequency warping will
easily submerge the main frequency component during data analysis, which will affect
the accuracy of signal and the results of subsequent data processing. Therefore, in
order to improve data quality and facilitate data processing, it is necessary to extract
or eliminate this change. The transfer learning fine-tuning formula and the CNN feature
extraction formula are shown in Eqs. (5) and (6):
Abnormal event of water pollution Water pollution refers to the change of chemical,
physical and biological characteristics of water samples caused by the inflow of certain
substances, which leads to the deterioration of water environment and affects the
function of water bodies [21]. The occurrence of abnormal water quality events often lasts for a long time, and
there are differences between the monitoring value and the expected value of water
quality indicators collected by sensors. Therefore, how to correctly detect the events
caused by water pollution is the core content of this paper.
2.3. The Process of Water Quality Anomaly Detection
Water quality anomaly detection primarily involves analyzing monitoring index data
obtained from relevant departments. Using soft-sensing technology, anomalies in water
quality are identified through data analysis methods. The process typically includes
assessing real-time monitoring data by preprocessing the collected data and comparing
it with predefined thresholds to determine the water quality status. Additionally,
residual sequence analysis involves a four-step process: Initially, a prediction model
is trained with historical time series data from the water quality monitoring department.
This model aims to forecast future fluctuations in the water quality index. Then selecting
the appropriate sample index as the input parameter to obtain the prediction value;
furthermore, comparing the prediction value output by the model with the actual value
obtained by the sensor to output the prediction residual sequence; finally, comparing
the residual sequence with the set threshold or further clustering the prediction
residual sequence, the water quality is divided into abnormal and normal categories.
Fig. 2. Step diagram of image processing optimization based on transfer learning.
Fig. 2 shows step diagram of image processing optimization based on transfer learning. Because
the water quality indexes to be detected belong to unlabeled time series data, this
paper studies the detection technology based on neural network model and unsupervised
clustering anomaly detection algorithm respectively [22]. The calculation formula of the identification accuracy of water quality parameters
and the calculation formula of water quality evaluation index are shown in formulas
(7) and (8).
3. Evaluation Method of Water Environment Quality Based on Improved Topis Method
In the comprehensive evaluation of water environment, the selection of evaluation
methods is the most important. Because of the uncertainty of water environment itself
and the diversity of various indexes (factors) affecting water environment, it is
particularly important to select appropriate water environment quality evaluation
methods and evaluate water environment quality reasonably and correctly [23,
24]. However, numerous water quality assessment methods studied and implemented both
domestically and internationally exhibit significant differences in theoretical aspects,
including evaluation procedures, standards, and classification principles. Each method
of water quality assessment has its own advantages and disadvantages, which determines
that the evaluator should choose according to the specific water environment pollution
type, assessment purpose, assessment standard, selection of assessment parameters
and sample size of monitoring data in the study area.
Fig. 3. Construction and verification process of water quality evaluation model.
Fig. 3 shows construction and verification process of water quality evaluation model. TOPSIS
method is a technique commonly used in multi-objective decision analysis of finite
schemes in system engineering in recent years, and has been used in the optimization
of pollution control schemes, etc. It is a method for giving attribute preference
information. In order to make TOPSIS method more suitable for evaluating water environment
quality, it is necessary to modify the shortcomings of traditional TOPSIS method and
form an improved TOPSIS method.
3.1. TOPSIS
3.1.1 Ideal solution
In this context, the ideal solution is a theoretical, hypothetical optimal scheme
that does not exist in reality. This scheme assumes that all attributes within the
decision matrix achieve their optimal values and is denoted as U. It represents the
ideal state that the decision-maker aims to approximate or strive towards.
3.1.2 Negative ideal solution
Negative ideal solution, contrary to ideal solution, represents the virtual worst
scheme, and when it is recorded, it represents the situation that is avoided as much
as possible.
3.1.3 Euclidean distance
The full name of Euclidean distance is Euclidean distance, which is a commonly used
definition of distance. It refers to the real distance between two points in m-dimensional
space. Euclidean distance can indicate the proximity between the evaluated scheme
and the best scheme and provide the quality of the evaluated scheme.
3.1.4 Relative proximity
The relative proximity degree is based on the Euclidean distance between the participating
scheme and the negative ideal solution and the ideal solution, which indicates the
proximity degree between the participating scheme and the optimal solution, and is
an important basis for determining the quality of the scheme [25].
3.2. Characteristics of TOPSIS
TOPSIS method does not require or limit the sample size of evaluation object, the
number of indicators and the distribution of data. It is not only suitable for data
with few programs and indicators and small sample size, but also suitable for data
with large programs and indicators and large sample size, which has good intuition
and is more flexible and simpler to use [26]. The water quality classification determination formula is shown in Eq. (9).
The traditional TOPSIS method generally uses the weighting method for the subjective
weighting method, such as expert scoring method, Delphi method and so on.
Fig. 4. Application flowchart of topsis algorithm combined with transfer learning
in water quality monitoring.
Fig. 4 shows application flowchart of TOPSIS algorithm combined with transfer learning in
water quality monitoring. Although it is widely used, it has some limitations and
the weighting results are not comprehensive enough. When calculating the distance
between each participating sample based on it, there are often Euclidean distances
of participating schemes that are close to both ideal solution and negative ideal
solution, which leads to the inability to fully reflect the quality of participating
schemes in ranking. The distance calculation formula in the TOPSIS algorithm is shown
in Eq. (10).
The traditional TOPSIS (Technique for Order Preference by Similarity to Ideal Solution)
method evaluates the proximity of a sample scheme to the ideal solution by calculating
the distance between each participating sample and both the ideal solution and the
negative ideal solution. However, there may be instances where a scheme exhibits close
proximity to both the Euclidean distance of the ideal solution and the distance of
the negative ideal solution [27]. Therefore, sorting the participating schemes simply according to Euclidean distance
sometimes cannot fully reflect the quality of each participating scheme. According
to the relevant theoretical research results, this paper improves the calculation
of index weight and relative proximity algorithm, and the improved TOPSIS method can
avoid various unreasonable results.
3.3. Evaluation Index and Data Source
Seven representative indicators—chemical oxygen demand (COD), BOD5, ammonia nitrogen,
volatile phenol, cadmium, petroleum, and total phosphorus—were selected as evaluation
factors. The data required for the evaluation were obtained through experiments conducted
at the Water Quality Monitoring Center of the Hydrology and Water Resources Bureau,
upstream of the Lanzhou Yellow River Conservancy Commission. [28].
Fig. 5. Data distribution diagram of TOPSIS algorithm in water quality resource monitoring.
Fig. 5 shows data distribution diagram of TOPSIS algorithm in water quality resource monitoring.
In 2013, ammonia nitrogen and BOD were significant in the dry season, COD and these
two parameters were crucial in the normal water period, while COD, total phosphorus,
and BOD were key in the wet season. Overall, ammonia nitrogen, COD, and BOD significantly
influenced water quality, reflecting the substantial impact of organic pollutants
and domestic sources on the aquatic environment. The formula utilized for analyzing
the water quality change trend is presented in Eq. (11).
Fig. 6 illustrates the comparison of water environment categories before and after image
processing using transfer learning techniques. This analysis considers the wet, normal,
and dry seasons. A comprehensive evaluation of the water quality categories across
three monitoring sections yields the following ranking: Xincheng Bridge predominantly
falls under Class I, followed by Zhongshan Bridge and Baolan Bridge, exhibiting a
gradual decline in water quality from Class I to Class V. A similar trend is observed
for the other two sets of comparisons, with Class I water quality being the most prevalent,
followed by varying degrees of water quality decline.
These results indicate that the water environment in this area is relatively healthy
and requires minimal intervention to maintain its current state. [29]. The C* values of the three sections in wet season are close to Class II, so the
water quality of the three sections in wet season is Class II; The C* value of Xincheng
Bridge and Zhongshan Bridge sections is close to Class II and greater than Class I
in normal water period, while the C* value of Baolan Bridge is smaller than Class
II, but closer to Class II, so the water quality of the three sections also belongs
to Class I in normal water period. In dry season, the water quality C* values of Xincheng
Bridge, Zhongshan Bridge and Baolan Bridge are closer to Class I, so their water quality
belongs to Class II. The calculation formula of the comprehensive water quality evaluation
index is shown in Eq. (12):
Fig. 6. Comparison diagram of before and after image processing based on transfer
learning.
Fig. 7. Analysis diagram of prediction accuracy of water quality evaluation model
and TOPSIS algorithm.
Based on the predefined water quality targets for each section along the Lanzhou segment
of the Yellow River, the evaluation results show that the water quality in each section
meets the respective class standards specified in the Surface Water Environmental
Quality Standard (GB3838-2002) across all periods. Notably, the Xincheng Bridge section,
located within the Lanzhou water source quasi-protected area, is designated to achieve
a water quality target of Class II. The evaluation confirms that the Xincheng Bridge
section meets this standard as well. However, while all sections may meet Class I
standards, the degree of pollution varies. The improved TOPSIS method effectively
highlights these differences by measuring the distance between each evaluation index’s
monitoring value and the ideal point.
Fig. 7 shows analysis diagram of the prediction accuracy of water quality evaluation model
and TOPSIS algorithm. From the comparison results, we can see that the performance
of anomaly detection method based on traditional clustering is poor, which is greatly
affected by noise and cannot effectively avoid the impact of noise. This is because
the traditional fuzzy clustering method has a good ability to discriminate unsupervised
water quality data, but the time series data without feature extraction and noise
reduction are complex, and it is impossible to accurately discriminate anomalies only
by mapping direct nonlinear water quality data to high-dimensional space to calculate
similarity features such as spatial distance [30]. At the same time, compared with other methods, it is found that after analyzing
the time-frequency characteristics of water quality signals and decomposing them into
signal screening components of different scales, the influence of noise is effectively
reduced in the reconstruction of water quality signals, and the false alarm rate in
the process of anomaly detection is generally reduced.
4. Conclusion and Prospect
On the basis of comprehensive analysis and comparison of various water quality evaluation
methods, this paper selects and constructs an improved TOPSIS comprehensive evaluation
model to carry out comprehensive evaluation of water quality. The improved TOPSIS
method offers a more refined approach to index weighting, effectively addressing the
challenge of the proximity between the monitoring values and both the ideal and negative
ideal points. Consequently, the evaluation results derived from the improved TOPSIS
method are more convincing than those of the traditional method, and it has been observed
that these results generally outperform those of the traditional TOPSIS method.
The improved TOPSIS method is relatively simple in principle, simple in operation
process and intuitive in results, and can reflect different degrees of pollution in
the same type of water quality, so it can be applied to water environment quality
assessment. Single factor evaluation method emphasizes the decisive role of the single
index with the most serious pollution on the comprehensive water quality, which often
makes the water quality level at a poor level. Although it is easy to attract attention,
too pessimistic water quality evaluation results are also not conducive to the rational
allocation of water resources.
This paper details the technical framework and theoretical system for water quality
evaluation in the study area, focusing on pollution source investigation, water quality
monitoring, and the determination of evaluation factor weights. The weights of water
quality indices for the Lanzhou section of the Yellow River are determined using both
subjective and objective methods, with separate considerations for different water
diversion periods. It was found that the weights of each index vary across water periods,
but chemical oxygen demand (COD) and BOD consistently have relatively high weights,
indicating their significant impact on water environment quality.
The evaluation results show that the water quality of three monitoring sections in
Lanzhou section of the Yellow River has reached the standard of Class II water, and
Xincheng Bridge section, as the representative section of Lanzhou surface water quasi-water
source protection area, has reached the standard of Class II water. The water quality
of each section is different. Through the analysis of the trend of pollutants along
the way and the changes in each water period, it can be seen that the water quality
of Baolanqiao section is worse than that of the other two sections, which is related
to the large amount of sewage and wastewater accepted by the river from upstream to
downstream. The concentration of most pollutants tends to be higher during the dry
season compared to the other two water periods, a trend that is attributed to the
seasonal variation in the water quantity of the Yellow River. Additionally, certain
pollutants, such as chemical oxygen demand, experience significant fluctuations during
the wet season, which is linked to the substantial changes in sediment levels during
this period. The reason is that sediment itself can adsorb and consume organic matter,
which has an impact on the concentration change of chemical oxygen demand in water.
Funding
Funding Information:2022 Natural Science Research Project of Anhui Provincial Department
of Education, entitled "Application and Research of Improved Combination Weighting-TOPSIS
Method in Water Quality Assessment"; Grant No.: 2022AH052434.
References
C. Chai , L. Wang , D. Chen , J. Zhou , N. Li , H. Liu , Quantifying
future water resource vulnerability in a high-mountain third pole river basin under
climate change, Journal of Environmental Management, Vol. 367, Art. no. 121954, 2024
Y. Lu , X. Yang , D. Bian , Y. Chen , Y. Li , Z. Yuan , K. Wang , A
novel approach for quantifying water resource spatial equilibrium based on the regional
evaluation, spatiotemporal heterogeneity and geodetector analysis integrated model,
Journal of Cleaner Production, Vol. 424, Art. no. 138791, 2023
M. Kasiselvanathan , C. V. S. R. Prasad , J. V. Arputharaj , A. Suresh ,
M. Sinduja , K. B. Prajna , M. Shanmugm , Prediction of ground water quality
in western regions of Tamilnadu using LSTM network, Groundwater for Sustainable Development,
Vol. 25, Art. no. 101156, 2024
S. Manikandan , S. R. Deena , R. Subbaiya , D. S. Vijayan , S. Vickram ,
B. Preethi , N. Karmegam , Waves of change: Electrochemical innovations for environmental
management and resource recovery from water–A review, Journal of Environmental Management,
Vol. 366, Art. no. 121879, 2024
S. S. Matsuzaki , A. Kohzu , M. Watanabe , N. I. Kondo , A. Tatsuta , Use
of legacy nitrogen as a resource: Unfertilized lotus fields contribute to water quality
improvement and biodiversity conservation, Nature-Based Solutions, Vol. 4, Art. no.
100080, 2023
M. S. Sajna , T. Elmakki , K. Schipper , S. Ihm , Y. Yoo , B. Park ,
H. Park , H. K. Shon , D. S. Han , Integrated seawater hub: A nexus of sustainable
water, energy, and resource generation, Desalination, Vol. 571, Art. no. 117065, 2024
A. Nanda , N. Das , G. Singh , R. Bindlish , K. M. Andreadis , S. Jayasinghe
, Harnessing SMAP satellite soil moisture product to optimize soil properties to
improve water resource management for agriculture, Agricultural Water Management,
Vol. 300, Art. no. 108918, 2024
A. Nativio , Z. Kapelan , J. P. van der Hoek , Risk assessment methods for
water resource recovery for the production of bio-composite materials: Literature
review and future research directions, Environmental Challenges, Vol. 9, Art. no.
100645, 2022
N. Rey-Martínez , A. Guisasola , J. A. Baeza , Assessment of the significance
of heavy metals, pesticides and other contaminants in recovered products from water
resource recovery facilities, Resources, Conservation and Recycling, Vol. 182, Art.
no. 106313, 2022
U. W. R. Siagian , L. Lustiyani , K. Khoiruddin , S. Ismadji , I. G. Wenten
, S. Adisasmito , From waste to resource: Membrane technology for effective treatment
and recovery of valuable elements from oilfield produced water, Environmental Pollution,
Vol. 340, Art. no. 122717, 2024
M. Sun , L. Zhang , R. Yang , X. Li , J. Zhao , Q. Liu , Water resource
dynamics and protection strategies for inland lakes: A case study of Hongjiannao Lake,
Journal of Environmental Management, Vol. 355, Art. no. 120462, 2024
H. Thuret-Benoist , V. Pallier , G. Feuillade-Cathalifaud , Monitoring of the
impact of the proliferations of cyanobacteria on the characteristics of natural organic
matter in a eutrophic water resource: Comparison between 2012-2013 and 2017-2018,
Chemosphere, Vol. 291, Art. no. 132834, 2022
M. G. Uddin , A. Rahman , F. R. Taghikhah , A. I. Olbert , Data-driven evolution
of water quality models: An in-depth investigation of innovative outlier detection
approaches–A case study of the Irish Water Quality Index (IEWQI) model, Water Research,
Vol. 255, Art. no. 121499, 2024
C. Wu , K. Tang , C. Lu , Y. Zhao , X. Zhang , Q. Sun , L. Yan , Resource
and environmental risk assessment of groundwater well fields in the Beijing-Tianjin-Hebei
region, Groundwater for Sustainable Development, Vol. 26, Art. no. 101235, 2024
M. Anand , S. K. Goswami , D. Chatterjee , A. Bhattacharya , M. J. Piran
, Modified Jaya optimization and TOPSIS for determining the optimal frequency in
LF-HVac, Applied Soft Computing, Vol. 158, Art. no. 111573, 2024
K. K. Chakravarthi , L. Shyamala , V. Vaidehi , TOPSIS-inspired cost-efficient
concurrent workflow scheduling algorithm in cloud, Journal of King Saud University-Computer
and Information Sciences, Vol. 34, No. 6, pp. 2359-2369, 2022
N. Ezhilarasan , C. Vijayalakshmi , Optimization of fuzzy programming with TOPSIS
algorithm, Procedia Computer Science, Vol. 172, pp. 473-479, 2020
G. Fu , B. Li , Y. Yang , C. Li , Re-ranking and TOPSIS-based ensemble feature
selection with multi-stage aggregation for text categorization, Pattern Recognition
Letters, Vol. 168, pp. 47-56, 2023
S. Fu , H. Fu , Modeling and TOPSIS-GRA algorithm for autonomous driving decision-making
under 5G-V2X infrastructure, Computers, Materials & Continua, Vol. 75, No. 1, pp.
1051-1071, 2023
D. N. Jayakumar , P. Venkatesh , Glowworm swarm optimization algorithm with TOPSIS
for solving multiple-objective environmental economic dispatch problem, Applied Soft
Computing, Vol. 23, pp. 375-386, 2014
J. Jin , H. Garg , Intuitionistic fuzzy three-way ranking-based TOPSIS approach
with a novel entropy measure and its application to medical treatment selection, Advances
in Engineering Software, Vol. 180, Art. no. 103459, 2023
J. Kacprzyk , A. Bozhenyuk , E. Gerasimenko , Lexicographic maximum dynamic
evacuation modeling with partial lane reversal based on hesitant fuzzy TOPSIS, Applied
Soft Computing, Vol. 144, Art. no. 110435, 2023
M. A. M. A. Kermani , A. Badiee , A. Aliahmadi , M. Ghazanfari , H. Kalantari
, Introducing a procedure for developing a novel centrality measure (sociability
centrality) for social networks using the TOPSIS method and a genetic algorithm, Computers
in Human Behavior, Vol. 56, pp. 295-305, 2016
R. A. Krohling , R. Lourenzutti , M. Campos , Ranking and comparing evolutionary
algorithms with Hellinger-TOPSIS, Applied Soft Computing, Vol. 37, pp. 217-226, 2015
R. A. Krohling , A. G. C. Pacheco , A-TOPSIS–An approach based on TOPSIS for
ranking evolutionary algorithms, Procedia Computer Science, Vol. 55, pp. 308-317,
2015
P. K. Kwok , H. Y. K. Lau , Hotel selection using a modified TOPSIS-based decision
support algorithm, Decision Support Systems, Vol. 120, pp. 95-105, 2019
X. Li , Y. Luo , H. Wang , J. Lin , B. Deng , Doctor selection based on
aspect-based sentiment analysis and the neutrosophic TOPSIS method, Engineering Applications
of Artificial Intelligence, Vol. 124, Art. no. 106599, 2023
G. B. do Nascimento , M. dos Santos , Performance evaluation of machine learning
algorithms for network anomaly detection: An approach through the AHP-TOPSIS-2N method,
Procedia Computer Science, Vol. 214, pp. 164-171, 2022
E. Roszkowska , T. Wachowicz , Application of fuzzy TOPSIS to scoring the negotiation
offers in ill-structured negotiation problems, European Journal of Operational Research,
Vol. 242, No. 3, pp. 920-932, 2015
R. Susmaga , I. Szczęch , D. Brzezinski , Towards explainable TOPSIS: Visual
insights into the effects of weights and aggregations on rankings, Applied Soft Computing,
Vol. 153, Art. no. 111279, 2024
Jingjing Lan obtained her B.E. degree in agricultural water conservancy engineering
from Changchun Institute of Technology in 2011. She received her M.E. degree in agricultural
water and soil engineering from Hohai University in 2014. Presently, she is working
as an associate professor in the School of Hydraulic Engineering, Wanjiang University
of Technology. Her areas of interest are water resources planning and management,
water environmental protection, and water-saving irrigation.
Peng Zhang received his bachelor’s degree in engineering from North China University
of Water Resources and Electric Power in 2011, and a doctorate degree in engineering
from HoHai University in 2022. He is currently working as an Associate Professor at
the School of Hydraulic Engineering of WanJiang University of Technology His research
areas and directions include water pollution control, water resources development,
and remediation of polluted environments.
Lingjun Wu obtained her bachelor of science degree in marine technology from Hohai
University in 2008, and her master’s degree in physical oceanography from the same
university in 2011. She is currently a Lecturer at the School of Hydraulic Engineering,
Wanjiang University of Technology. Her research interests include port engineering
optimization and concrete durability.
Shuangshuang Huang graduated from Anhui Agricultural University with a bachelor of
engineering degree in agricultural water conservancy engineering in 2014. She obtained
a master of engineering degree in agricultural soil and water conservation engineering
from Hohai University in 2017. Currently, she serves as a lecturer at the School of
Hydraulic Engineering of Wanjiang Institute of Technology. Her research interests
include water resources planning and management, crop nutrient management, and water-saving
irrigation.
Jinxi Zhang holds his bachelor of engineering in Hydrology and Water Resources Engineering
from Changchun Institute of Technology (2017) and a master of Engineering in Agricultural
Engineering from Shihezi University (2020). Currently he is a lecturer at the School
of Hydraulic Engineering, WanJiang University of Technology. His research focuses
on medium-and long-term water resources planning and design, as well as agricultural
ecological environment regulation.