Energy Reports 8 (2022) 9119–9126

Contents lists available at ScienceDirect

Energy Reports
journal homepage: www.elsevier.com/locate/egyr

Optimal allocation of regional water resources based on simulated

annealing particle swarm optimization algorithm
∗
Zhanping Wang a,b , Juncang Tian a,c,d , , Kepeng Feng a,c,d
a
College of Civil and Hydraulic Engineering, Ningxia University, Yinchuan 750021, China
b
School of Mathematics and Statistics, Ningxia University, Yinchuan 750021, China
c
Ningxia Research Center of Technology on Water-Saving Irrigation and Water Resources Regulation, Yinchuan 750021, China
d
Engineering Research Center for the Efficient Utilization of Water Resources in Modern Agricultural in Arid Regions, Yinchuan 750021, China

article info a b s t r a c t

Article history: Given the fast growth of economy in Ningxia, China, contradictions between the increase of water
Received 27 April 2022 resources and the decrease of water supply become increasingly prominent. Therefore, it is critical to
Received in revised form 19 June 2022 utilize limited water resources in a rational manner. This study uses the multi-objective programming
Accepted 8 July 2022
theory and forms a multi-objective optimum allocation model for the purpose of using regional water
Available online 23 July 2022
resources sustainably. By aiming at maximizing the comprehensive benefits of society, the model
Keywords: solves the problem that particles are prone to be trapped into local minima by introducing the idea
Water resources of simulated annealing into the basic particle swarm optimization (PSO) algorithm. The simulated
Optimal allocation annealing (SA) particle swarm algorithm is applied to solve the model and obtain the optimum
Simulated annealing allocation schemes of water resources at three different precipitation frequencies in the planning year
Particle swarm algorithm of Yinchuan (2025), the capital city of Ningxia, China. With this, the model provides a scientific basis
for the management of water resources in the city. The results indicate that the model is built upon
a scientific and practical foundation, and the algorithm has practical significances.
© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Contents

1. Introduction..................................................................................................................................................................................................................... 9119
2. Methodology ................................................................................................................................................................................................................... 9120
2.1. Optimal water resource allocation model ...................................................................................................................................................... 9120
2.1.1. Objective function .............................................................................................................................................................................. 9120
2.1.2. Constraints .......................................................................................................................................................................................... 9121
2.2. Simulated annealing particle swarm algorithm ............................................................................................................................................. 9121
2.2.1. Simulated annealing algorithm ....................................................................................................................................................... 9121
2.2.2. Particle swarm optimization algorithm........................................................................................................................................... 9121
2.2.3. Simulated annealing particle swarm optimization algorithm ...................................................................................................... 9122
3. A case study in Yinchuan .............................................................................................................................................................................................. 9122
3.1. overviews of Yinchuan City .............................................................................................................................................................................. 9122
3.2. Data sources ....................................................................................................................................................................................................... 9122
3.3. Model parameters.............................................................................................................................................................................................. 9122
3.4. Algorithm parameters ....................................................................................................................................................................................... 9123
4. Results analysis............................................................................................................................................................................................................... 9124
5. Conclusion ....................................................................................................................................................................................................................... 9125
Declaration of competing interest................................................................................................................................................................................ 9125
Data availability .............................................................................................................................................................................................................. 9125
Acknowledgments .......................................................................................................................................................................................................... 9125
References ....................................................................................................................................................................................................................... 9126

1. Introduction
∗ Corresponding author at: College of Civil and Hydraulic Engineering, Ningxia
University, Yinchuan 750021, China. With the persistent progress of society and the fast devel-
E-mail address: [email protected] (J. Tian). opmental pace of economy, human beings are exploiting and
https://doi.org/10.1016/j.egyr.2022.07.033
2352-4847/© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

utilizing limited water resources unprecedentedly, so that the algorithms cannot acquire satisfactory allocation results due to
water resource shortage has evolved from a regional problem in the limitations of themselves. Therefore, in the process of the
certain water-deficient nations to a challenge across the globe OAWRs, it is imperative to select an appropriate model, to over-
(Bagatin et al., 2014; Schwarzenbach et al., 2010). The conflict be- come the shortcomings of the algorithms, and to ameliorate the
tween the demand and supply of water resources becomes more convergent accuracy and efficiency of the arithmetic. Hence, in
and more serious, and problems such as water shortage and water the paper, an in-depth analysis was carried out on the PSO and SA
environment deterioration have become the restrictive factors arithmetics. On condition that the characteristics of some inferior
for social development. Failure to achieve equitable distribution solutions were accepted temporarily, the SA algorithm controlled
will lead to conflict among water sectors (Roozbahani et al., at a certain possibility can be used to ameliorate the local opti-
2020). In the context where water resources problem continues mization capability of the basic SPO arithmetic to obtain a SA -
deteriorating, the optimal allocation of water resources (OAWRs) PSO algorithm with a better global optimization ability, which is
has important theoretical significance and application value. expected to provide a new calculating method for future water
Many scholars domestically and internationally have com- resources planning models and provide more decision-making
pleted a great deal of researches on the establishment of models basis for regional water resources managers.
in the OAWRs. OAWRs helps to establish an equilibrium between
water supply and demand (Roozbahani et al., 2020; Tarebari 2. Methodology
et al., 2018). To alleviate water scarcity and conflict, a variety of
different mathematical programming methods have been devel- 2.1. Optimal water resource allocation model
oped (Tarebari et al., 2018). Single-objective and multi-objective
models are the methods currently used to allocate water re- The OAWR model is aimed at protecting the ecosystem, achiev-
sources (Dai et al., 2018; Davijani et al., 2016; Fu et al., 2018;
ing the overall goal of regional socioeconomic sustainability, and
Hu et al., 2016a,b). Globally, computer technologies are used
boosting socioeconomic sustainability. The three objectives cover
to simulate the complex relationship of water resources sys-
economy, society and ecological environment, and the objective
tems, so as to form an allocation method widely accepted (Hao
function is stated below:
and Jinjun, 2008). Masse (1990) researched the single-objective
problem in water resource allocation, and the results are fea- F (X) = max {f1 (x) , f2 (x) , f3 (x)} . (1)
tured by applying the allocation theory into practice. Abdulbaki
et al. (2017) proposed and solved the multi-objective problem Where, f1 (x) represents the economic goal, f2 (x) represents the
in water resources allocation. An allocation model of regional societal goal, and f3 (x) denotes the environmental goals. There
multi-water sources and multi-user water resources with differ- are lots of influencing factors in the water resources system,
ent water quality was established by Wada et al. (2014), and and those factors have different influence scales and depths on a
a multi-source joint scheduling model was established by El- target while restricting each other. Therefore, the problem related
naqa et al. (2007) based on the comprehensive consideration to the OAWRs is multiple objectives.
of groundwater overexploitation and water quality deteriora-
tion. Rosegrant et al. (2000) proposed the use of a coupled eco- 2.1.1. Objective function
nomic model and a hydrological model in research, which were Three objective functions are established respectively for the
applied in the watershed development planning in Mexico. An 3 goals of regional OAWR, including economy, society and envi-
optimal allocation model for water resources proposed by Clau- ronment goals, which are mutually coordinated.
dia (2002) can be applied to determine a reasonable allocation Suppose the study area had K sub-areas, and k = 1, 2, . . . , K ;
scheme by rating and analyzing the value of water resources. To there were I(k) independent water sources (IWSs) and J(k) users
fulfill the requirements of reservoir optimization and schedul- in the k sub-areas; M public water resources were in the research
ing, Zhong et al. (2003) proposed a multi-objective decomposition area, and.
model related to the optimal scheduling of water resources and (1) Economic objective
found a solution by mathematical simulation method. Ma (2004) The maximal direct economy benefits generated by the local
analyzed the situation related to the reuse of water resources in water supply is taken as the economic objective, and the objective
detail, established a fuzzy linear optimized allocation model and function is:
allocated water resources based on this model. It helps address
the problem of insufficient water irrigation in agriculture. Li et al. K J(k) I(k)
(2009) put forward a fuzzy multi-stage random programming
∑ ∑ ∑(
bkij − cijk xkij αik βjk ωk
)
maxf1 (X ) =
modeling method in the research and described its application in
k=1 j=1 i=1
detail. Lv et al. (2010) established a corresponding basin-based
K J(k) M
water resources management model and found a solution to ∑ ∑ ∑(
bkcj − ccjk xkcj αck βjk ωk .
)
the fuzziness of relevant interval values using this model. Zhang + (2)
and Li (2014) established interval parameters and a stochastic k=1 j=1 c =1

programming model in research and used the model to solve Where, xkij and xkcj are decisive variates, which are the water
the problems related to the sustainability of water resources
supply volume (10,000 m3 ) from the IWS i and the public water
in Heilongjiang. A solution to fuzzy programming was intro-
source (PWS) c to the user j in the sub-area k respectively; bkij ,
duced by Zhang et al. (2015) for planning regional sustainable
development. Wang et al. (2016) established a 2-stage random and bkcj are the benefit coefficients (yuan/m3 ) of water supply; cijk
programming model taking into account the value-at-risk theory and ccjk denote the cost coefficients (yuan/m3 ); αik and αck denote
in research and used the model to solve problems of different the water supply order coefficients of the IWS i and PWS c in
allocation types, thus providing support for the improvement of the subarea k; βjk denotes the water using equity coefficient of
the resource utilization level. the user j in the subarea k; ωk is the weight coefficient of the k
The number of optimal allocation models of water resources sub-area.
grow exponentially due to their own diversification, so simple (2) Social goal
linear programming algorithms can no longer meet the devel- The measurement of the social objective is not as easy as that
opment trend of the models. Even some intelligent optimization of the economy goal. Herein, the societal goal is to minimize
9120
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

the water shortage in the regional water supply system, and the Where, µ and µ∗ denote the local coordination development
objective function is as follows: indicator and its minimal value. µB1 (σ1 ) and µB2 (σ2 ) represent the
K J(k)
[ ( I(k) M
)] coordination development level between local water resource
usage and local economy achievement and the coordination de-
∑ ∑ ∑ ∑
min f2 (X ) = Dkj − xkij + xkcj . (3)
velopment level between local economy achievement and envi-
k=1 j=1 i=1 c =1
ronmental quality amelioration, respectively.
In which, Dkj denotes the water demand of user j in sub-area k (5) Variable non-negative constraints
(10,000 m3 ).
xkij , xkcj ≥ 0. (11)
(3) Environment goal
The ecological environment goal is to minimize the discharge
of important regional pollutant while ensuring the ecological 2.2. Simulated annealing particle swarm algorithm
environment water supply. The objective function is:
2.2.1. Simulated annealing algorithm
J(k)
K
( I(k) M
)
∑ ∑ ∑ ∑ The SA arithmetic is a random global optimization arithmetic
min f3 (X ) = 0.01dkj pkj xkij + xkcj . (4) proposed by Metropolis et al. (1953) on the foundation of the re-
k=1 j=1 i=1 c =1 semblance between the annealing process of solid matter and the
In which, dkj denotes the content (mg/L) of important pollution general process of solving combinational optimization problems.
factors (COD and BOD) per unit wastewater discharge volume The steps are as follows.
of the user j in the subarea k; pkj is the wastewater discharge Step 1: set an adequately large original temperature T0 > 0;
coefficient of the user j in the subarea k. annealing times k = 0; let, with initializing, and calculate its
energy value.
2.1.2. Constraints Step 2: confirm whether the set cycle termination condition is
(1) Constraints on the water supply capability of the regional met at the temperature; if so, go to step 3;
( otherwise, randomly
select a new solution x(j) in the field N( x(i)) of the
)
water supply system point, and
calculate the energy difference ∆Eij = E x((i) − E) x(J ) between
( )
Constraints on the water supply capability of the PWS:
∆Eij
⎧ J(k) them; if, let x(i) = x(j) ; if, solve p = exp − Tk
; if p > η [a
⎪ ∑
xkcj ≤ W (c , k) , random number distributing randomly and uniformly within (0,
⎪
⎪
⎪
⎪
⎨
j=1 1)], let, otherwise, repeat step 2.
(5) Step 3: solve Tk+1 = d (Tk ) according to a annealing equation
⎪ K
d(Tk ); confirm whether the Metropolis criterion is met; if so, the
∑
W (c , k) ≤ Wc .
⎪
⎪
⎪
⎪
⎩ calculation is over, and then output the result, otherwise, go back
k=1
to step 2.
Where, W (c , k) denotes the water volume allocated by the
PWS c to the subarea k; Wc denotes the available water volume
2.2.2. Particle swarm optimization algorithm
of the PWS c.
The SPO arithmetic is an efficient search algorithm with high
Constraints on the water supply capacity of the independent
feasibility proposed by Kennedy and Eberhart (1995) after being
water source:
illuminated by the foraging behaviors of birds. It belongs to the
J(k)
∑ evolutionary algorithm like the simulated annealing algorithm,
xkij ≤ Wik . (6) and both search for the optimum solution iteratively. The steps
j=1 are as follows:
Step1: initialize a particle population with the population size
Where, Wik denotes the available water volume of the IWS i in
N, the initial position popi and the velocity.
subarea k.
Step2: compute the fitness value fitness(i) of every particulate.
(2) Constraints on water delivery capability
Constraints on the water delivery capability of the IWS: Step 3: contrast the individual extreme value pbest (i) with
the fitness value fitness(i) of every particulate, if, let pbest (i) =
xkij ≤ Pijk . (7) fitness (i).
Step 4: contrast the global extreme value gbest (i) with the
Where, Pijk denotes the maximal water delivery capability of fitness value fitness(i) of every particulate, if fitness(i) > gbest (i),
the IWS i in the k subarea to the j user. let gbest (i) = fitness (i).
Constraints on the water delivery capability of the PWS: Step 5: update the particulate’s position popi and velocity.
W (C , K ) ≤ Pck . (8)
vid (t + 1) = wvid (t )+c1 r1 [pid − popid (t)]+c2 r21 pgd − popid (t) ,
[ ]
In which, Pck denotes the maximal water delivery capability of
(12)
the PWS c to the k sub-area.
(3) Constraints on the water demand capacity of the water
system popid (t + 1) = popid (t ) + vid (t + 1) . (13)
I (k)
∑ M
∑ In which, w denotes the inertia weight; c1 and c2 are learning
Lkj ≤ xkij + xkcj ≤ Hjk . (9) factors, representing the individual cognition and social acceler-
i=1 c =1 ation constant of i separately; r1 and r2 are stochastic numbers
independent of each other in [0,1].
In which, and Hjk are the upper and lower limits of the water
Lkj
Step 6: confirm if the terminate condition is met; if the result
demand of the user j in the subarea k, respectively.
meets the precision requirements or reaches the maximal quan-
(4) Constraints on regional coordinated development
tity of the iterative process, the calculation is over, otherwise, go
√
µ= µB1 (σ1 ) µB2 (σ2 ) ≥ µ∗ . (10) back to step 2.
9121
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

Table 1 depth of 187 mm, with an increase of 9% over the average annual
Water demand of Yinchuan in 2015.
rate of past years. The amount of natural surface water resources
Zone Agriculture Industry Domesticity Ecology Total
is 97.5 million m3 , and the amount of groundwater resources is
Yinchuan 39839 6012 5552 2705 54108 504.4 million m3 . The double counting amount between them
Yongning 43023 688 402 396 44508
Helan 51050 633 332 941 52956
is 444.5 million m3 . The total amount of the water resources in
Lingwu 26848 2954 484 19 30306 Yinchuan are 157.4 million m3 , as show in Fig. 2.
Total 160760 10287 6770 4060 181877

(Unit: ten thousand m3 ). 3.2. Data sources

Considering the rationality of allocation, water consumption

2.2.3. Simulated annealing particle swarm optimization algorithm in all industries is shown in Table 1 with 2015 as the base
The PSO arithmetic can effectively optimize various functions, year. Optimal allocation of water resources is carried out for
but they are prone to be trapped into local minima in solv- the three different scenarios of rainfall frequency of 50%, 75%
ing discrete functions and multi-objective optimization. The SA and 90% respectively in 2025, the target year of the program.
algorithm is capable of jumping out of the initial status and Based on the statistics of Yinchuan’s external water transfer data
persistently searching by accepting certain inferior solutions at
and water conservancy development plan over the past years,
a certain possibility. In the paper, the SA-PSO algorithm was
combined with the design purpose, water supply object, project
constructed by combining them two in the process of calculation,
layout and adjustment scheme, the available water supply of
so that the arithmetic had a potent global optimization capability
Yinchuan’s reservoir dam project, water transfer project and new
during searching and an ability of efficiently avoiding the search
planned project, water supply at different rainfall frequencies in
being trapped into local minima. The calculation steps are as
follows. the target year of the program. According to the Yinchuan Sta-
Step1: initialize a particle population with the population size tistical Yearbook 1995–2015, Water Resources Bulletin and other
N, the initial location popi and the velocity. relevant materials, and based on the regional socio-economic
Step2: compute the fitness value fitness(i) of every particulate; development indicators, the quota method and trend method in
contrast the individual extreme value pbest (i) with the fitness water resources planning are used for prediction, so as to obtain
value fitness(i) of every particulate, if, let pbest (i) = fitness (i); the water demand of different industries that need water in 2025,
likewise, contrast the global extreme value gbest (i) with the fit- as shown in Table 2.
ness value fitness(i) of every particulate, if fitness(i) > gbest (i), let
gbest (i) = fitness (i). 3.3. Model parameters
Step 3: set an adequately large original temperature T0 > 0
and carry out annealing accordingly, that is
The above models should be utilized for the OAWRs in Yinch-
T0 = f (pi )/ln5, (14) uan by following the principle of giving priority to domestic
and ecological water over the industrial and agricultural water.
Tk+1 = λTk . (15)
When determining the water supply order coefficient, the priority
In which, λ denotes the attenuation parameter in the course of degree of each water source is first converted into the coefficient
annealing. on the interval [0,1], and the specific calculation formula is as
Step 4: identify the fitness value of every particulate pi at the follows
current temperature T, that is

exp−(f (pi )−f (pg ))/T 1 + nk − nki

TF (pi ) = ∑M . (16) αik = ∑I(k) [ max ].
−(f (pi )−f (pg ))/T k k
i=1 exp j=1 1 + nmax − nj

In which, TF denotes the particulate fitness value. In the formula nki indicates the water i supply sequence number of
Step 5: identify its global optimum substitute value p′i from sub-districts k; nkmax indicates the maximum value of water supply
the entire, and update its location and speed by Eqs. (9) and (10).
sequence number in a sub-area k.
Step 6: determine the particulate target value by calculation,
When optimizing the allocation of water resources in Yinchuan,
and update pbest (i) and gbest (i) before annealing.
according to the characteristics of different users, following the
Step 7: confirm if the terminate condition is satisfied; if so, the
principle of ensuring domestic water first, water production sec-
calculation is over, and output the result; otherwise, go back to
ond, the water supply sequence is determined as domestic water,
step 4.
ecological water, industrial water, agricultural water. According
The flowchart of the arithmetic is presented by Fig. 1.
to the above formula, the water supply order coefficients αik of
3. A case study in Yinchuan surface water, ground water and water from the Yellow River
were 0.5, 0.33 and 0.17. The user equity coefficients βjk refers
3.1. overviews of Yinchuan City to the degree of preferential access to water supply for users j
in the sub-area k compared to other departments that also need
Yinchuan City is located in the middle of the Ningxia Plain in water, and the determination method of βjk is similar to αik , the
the upper reaches of the Yellow River, with an overall area of user equity coefficients of each user is obtained as follows: 0.4,
9555.38 square kilometers. Yinchuan and its subordinate counties 0.3, 0.2, 0.1. The weight coefficients of the three targets were
(cities) have a representative mid-temperate continental climate, determined by the binary comparison method to be 0.4, 0.33
with a yearly mean temperature of about 8.5 ◦ C, average annual and 0.27 respectively; the weight coefficients of sub-areas were
sunshine hours varying from 2,800 to 3,000, and an annual aver- 0.34, 0.20, 0.20 and 0.26, respectively. Water consumption de-
age precipitation of about 200 mm. In 2015, the total amount of partments’ efficiency coefficient and cost coefficient in Yinchuan
precipitation was 1.4068 billion m3 , equivalent to a precipitation in 2025, see Table 3.
9122
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

Fig. 1. Flowchart of SA-PSO arithmetic.

Table 2
The upper and lower limits of water consumption departments’ water demand in Yinchuan in 2025.
Zone Agriculture Industry Domesticity Ecology
Upper limit Lower limit Upper limit Lower limit Upper limit Lower limit Upper limit Lower limit
Yinchuan 70351 37236 16768 13355 8978 8978 4678 4678
Yongning 72293 43253 3534 2789 906 906 706 706
Helan 93768 58669 3809 3231 956 956 1768 1768
Lingwu 33686 31153 23936 19167 1123 1123 56 56

(Unit: ten thousand m3 ).

3.4. Algorithm parameters this paper set the particle population number N at 1000, inertia
weight ω 0.7886, learning factor c1 = c2 = 2, and r1 , r2 are
Using Matlab R2018b as the development tool to write the random numbers on the interval [0,1], the initial temperature
simulated annealing particle swarm optimization program. Ac- T0 of simulated annealing set at 1000, the termination temper-
cording to the previous experience, after several trial calculations, ature Tf 1, the attenuation parameter λ 0.7, and the maximum
9123
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

Fig. 2. Location of the case region.

Table 3
Water consumption departments’ efficiency coefficient and cost coefficient in Yinchuan in 2025.
Zone Efficiency coefficient Cost coefficient
Agriculture Industry Domesticity Ecology Agriculture Industry Domesticity Ecology
Yinchuan 2 588 500 500 0.0305 2.60 2.31 0.0305
Yongning 1.4 556 500 500 0.0305 2.90 2.08 0.0305
Helan 1.3 435 500 500 0.0305 2.90 2.00 0.0305
Lingwu 1.7 227 500 500 0.0305 2.40 2.00 0.0305

(Unit: yuan/m3 ).

Table 4
Results of optimal allocation of water resources at the precipitation frequency of 50% in Yinchuan in 2025.
Zone Item Agriculture Industry Domesticity Ecology Total
Water demand 34945 7194 11724 4253 58116
Yinchuan Water supply 33290 7002 11724 4253 56268
Water shortage 1656 192 0 0 1847
Water demand 36807 1240 1045 622 39715
Yongning Water supply 35040 1132 1045 622 37840
Water shortage 1767 108 0 0 1875
Water demand 44185 1141 865 1479 47670
Helan Water supply 42653 1064 865 1479 46061
Water shortage 1532 76 0 0 1608
Water demand 23680 2959 1261 30 27930
Lingwu Water supply 22750 2811 1261 30 26852
Water shortage 931 147 0 0 1078
Water demand 139618 12533 14895 6384 173430
Total Water supply 133733 12010 14895 6384 167022
Water shortage 5885 523 0 0 6408

(Unit: ten thousand m3 ).

iteration max Tit 1000. The model parameters were set up as per At the precipitation frequency of 75%, the total water demand
the supply and demand of water at three different precipitation of Yinchuan is 1,843.38 million m3 , with the water shortage of
frequencies (50%, 75%, 90%) in the programming year of Yinchuan 163.88 million m3 and the water shortage rate of 8.9%. Domestic
in 2025. The results of OAWRs in Yinchuan can be obtained by water is sufficient, but the water shortage of agriculture, industry
program-optimized calculation, see Tables 4, 5 and 6. and ecology is 148.55 million m3 , 11.15 million m3 and 4.18
million m3 , with the shortage rates of 9.9%, 8.9% and 6.5% respec-
4. Results analysis tively. Among them, Helan county faces the most serious water
shortage. Agriculture water shortage is more serious than any
Based on the water supply and demand in Yinchuan in 2025, other industries.
the SA-PSO arithmetic was utilized to establish the optimum At the precipitation frequency of 90%, the total water demand
allocation model, and the results of OAWR at three precipitation of Yinchuan is 1,974.27 million m3 , with the water shortage of
frequencies of 50%, 75%, and 90% in 2025 were obtained, as shown 361.76 million m3 and the shortage rates of 18.3%. The water
in Tables 4, 5 and 6. shortage is severe. Only the domestic water is sufficient. The wa-
The results indicate that at 50% precipitation frequency in ter shortage for agriculture, industry and ecology is 327.89 mil-
2025, the total water demand of Yinchuan will be 1,734.3 million lion m3 , 23.64 million m3 and 10.23 million m3 , with the shortage
m3 , with the shortage rate of 3.7%. Domestic and ecological water rates of 20.0%, 18.9% and 16.0% respectively. Among them, Helan
is sufficient, but agricultural and industrial water in shortage, county faces the most serious water shortage. Agriculture water
indicating that the water supply can basically satisfy the demand. shortage is more serious than any other industries.
9124
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

Table 5
Results of optimal allocation of water resources at the precipitation frequency of 75% in Yinchuan in 2025.
Zone Item Agriculture Industry Domesticity Ecology Total
Water demand 37675 7194 11724 4253 60846
Yinchuan Water supply 33772 6750 11724 4089 56335
Water shortage 3903 444 0 164 4511
Water demand 39683 1240 1045 622 42590
Yongning Water supply 35548 1098 1045 567 38259
Water shortage 4134 142 0 55 4332
Water demand 47637 1141 865 1479 51122
Helan Water supply 43272 960 865 1289 46385
Water shortage 4366 181 0 190 4737
Water demand 25530 2959 1261 30 29780
Lingwu Water supply 23079 2610 1261 21 26971
Water shortage 2451 349 0 9 2809
Water demand 150526 12533 14895 6384 184338
Total Water supply 135671 11418 14895 5966 167950
Water shortage 14855 1115 0 418 16388

(Unit: ten thousand m3 ).

Table 6
Results of optimal allocation of water resources at the precipitation frequency of 90% in Yinchuan in 2025.
Zone Item Agriculture Industry Domesticity Ecology Total
Water demand 40952 7194 11724 4253 64122
Yinchuan Water supply 32566 6210 11724 3989 54489
Water shortage 8386 984 0 264 9633
Water demand 43133 1240 1045 622 46041
Yongning Water supply 34279 760 1045 356 36440
Water shortage 8855 480 0 266 9601
Water demand 51780 1141 865 1479 55264
Helan Water supply 41726 889 865 998 44478
Water shortage 10054 252 0 481 10787
Water demand 27750 2959 1261 30 32000
Lingwu Water supply 22255 2310 1261 18 25844
Water shortage 5495 649 0 12 6156
Water demand 163615 12533 14895 6384 197427
Total Water supply 130826 10169 14895 5361 161251
Water shortage 32789 2364 0 1023 36176

(Unit: ten thousand m3 ).

5. Conclusion of water use are also taken into account. A satisfactory solution
will be obtained if it is applied to the optimal allocation of water
In the present research, an optimal model of a complex water resources in Yinchuan.
resource system with multiple water sources and multiple users
was established on the basis of taking into account ecological Declaration of competing interest
environment factors. The model functions as a decision support
system, which can allocate optimal water resources to each user.
The authors declare that they have no known competing finan-
The model embraces society, economy, and environment factors,
and the SA - PSO algorithm is utilized to solve allocation. cial interests or personal relationships that could have appeared
The PSO arithmetic can effectively optimize various functions, to influence the work reported in this paper.
but they are prone to be trapped into local minima in solv-
ing discrete functions and multi-objective optimization. The SA Data availability
algorithm is capable of jumping out of the initial status and
persistently searching by accepting certain inferior solutions at The data is already in the article
a certain possibility. In the paper, the SA-PSO algorithm was
constructed by combining them two in the process of calculation,
Acknowledgments
so that the arithmetic had a potent global optimization capability
during searching and an ability of efficiently avoiding the search
being trapped into local minima. On condition that the character- The research has obtained the support from the Priority Re-
istics of some inferior solutions were accepted temporarily, the SA search and Projects for Ningxia in China (2020BEG03021), the Sci-
algorithm controlled at a certain possibility was used to amelio- ence and Technique Research Foundation of Ningxia Institutions
rate the local optimization capability of the basic SPO algorithm, of Higher Education, China (No. NGY2020010), Natural Science
effectively overcomes the shortcomings of only using the particle Foundation of Ningxia Province (2020AAC03058, 2021AAC02007),
swarm algorithm or simulated annealing algorithm and avoid the First-class Major Foundation of Ningxia Institutions of High
particles from being trapped into local minima, thereby improv- Education in China (NXYLXK2021A03), the First-class Discipline
ing the accuracy of the algorithm and accelerating its evolution. Construction Project in Ningxia Universities: Mathematics, and
This method provides a better solution for the optimum allocation the Scientific Research Project of Ningxia Colleges and Univer-
of complex water resource systems, and the quantity and priority sities, China (NGY2020006).
9125
Z. Wang, J. Tian and K. Feng Energy Reports 8 (2022) 9119–9126

References Lv, Y., Huang, G.H., Li, Y.P., et al., 2010. Planning regional water resources system
using an interval fuzzy bi-level programming method. J. Environ. Inf. 16 (2),
Abdulbaki, D., Ahindi, M., Yassine, A., et al., 2017. Anoptimization model for the 43–56.
allocation of water resources. J. Clean. Prod. 164, 994–1006. Ma, W.-F., 2004. Fuzzy optimal allocation of reclaimed and natural water
Bagatin, R., Klemes, J.J., Reverberi, A.P., Huising, D., 2014. Conservation and resources for agricultural irrigation. J. Agro-Environ. Sci. 23 (4), 770–773,
improvements in water resource management: a global challenge. J. Clean. (In chinese).
Prod. 77, 1–9. Masse, A.K., 1990. Withdrawal of shelf water into an estuary:a barotropic model.
Claudia, Ringler, 2002. Optimal Allocation and Use of Water Resources in the J. Geophys. Res. 95 (C9), 16085–16096.
Mekong River Basin; Multi-Country and Intersectoral Analyses. Scitech Book Metropolis, N., Rosenaluth, A.W., Rosenbluth, M.N., et al., 1953. Equation of
News, Portland. state calculations by fast computing machines. J. Chem. Phys. 21 (6),
Dai, C., Qin, X., Chen, Y., Guo, H., 2018. Dealing with equality and benefit for 1087–1092.
water allocation in a lake watershed: a gini-coefficient based stochastic opti- Roozbahani, R., Abbasi, B., Schreider, S., Hosseinifard, Z., 2020. A basin-wide
mization approach. J. Hydrol. 561, 322–334. approach for water allocation and dams location–allocation. Ann. Oper. Res.
Davijani, M.H., Banihabib, M., Anvar, A.N., Hashemi, S., 2016. Multi-objective 287, 323–349.
optimization model for the allocation of water resources in arid regions Rosegrant, M.W., Ringler, C., D.C., Mckinny., et al., 2000. Integrated economic
based on the maximization of socioeconomic efficiency. Water Resour. hydrologic water modeling at the basin scale: The maipo river basin. Agric.
Manag. 30, 927–946. Econ. 24 (1), 33–46.
Elnaqa, A., Almomani, M., Kilani, S., et al., 2007. Groundwater deterioration of Schwarzenbach, R.P., Egli, T., Hofstetter, T.B., Gunten, U.von., Wehrli, B., 2010.
shallow groundwateraquifers due to overexploitation in Northeast Jordan.
Global water pollution and human health. Annu. Rev. Environ. Resour. 35,
Clean 35 (2), 156–166.
109–136.
Fu, Y., Zhang, J., Zhang, C., Zang, W., Guo, W., Qian, Z., et al., 2018. Payments for
Tarebari, H., Javid, A.H., Mirbagheri, S.A., Fahmi, H., 2018. Multi-objective surface
ecosystem services for watershed water resource allocations. J. Hydrol. 556,
water resource management considering conflict resolution and utility func-
689–700.
tion optimization. Water Resour. Manag. 32, 4487–4509.
Hao, Wang, Jinjun, You, 2008. Advancements and development course of
Wada, Y., Wisser, D., Bierkensm, F.P., 2014. Globalmodeling of withdrawal,
research on water resources deployment. J. Hydraul. Eng. 39 (10),
1168–1175. allocation and consumptive useof surface water and groundwater resources.
Hu, Z., Chen, Y., Yao, L., Wei, C., Li, C., 2016a. Optimal allocation of regional water Earthsystem Dyn. 5 (1), 15–40.
resources: from a perspective of equity–efficiency tradeoff. Resour. Conserv. Wang, Y., Huang, G.H., Wang, S., et al., 2016. A stochastic programming with
Recycl. 109, 102–113. imprecise probabilities model for planning water resources systems under
Hu, Z., Wei, C., Yao, L., Li, L., Li, C., 2016b. A multi-objective optimization model multiple uncertainties. Stoch. Environ. Res. Risk Assess. 30 (8), 2169–2178.
with conditional value-at-risk constraints for water allocation equality. J. Zhang, L., Li, C.Y., 2014. An inexact two-stage water resources allocation model
Hydrol. 542, 330–342. for sustainable development and management under uncertainty. Water
Kennedy, J., Eberhart, R., 1995. A new optimizer using particle swarm theory. Resour. Manag. 28 (10), 3161–3178.
In: Proceedings of the Sixth International Symposium on Micro Machine and Zhang, K., Li, Y.P., Huang, G.H., et al., 2015. Modeling for regional ecosystem
Human Science. IEEE, Nagoya, Japan. Piscataway, pp. 39–43. sustainable development under uncertainty—A case study of dongying, China.
Li, Y.P., Huang, G.H., Huang, Y.F., et al., 2009. A multistage fuzzy-stochastic Sci. Total Environ. 533, 462–475.
programming model for supporting sustainable water-resources allocation Zhong, P.-A., J.-N., Liu., et al., 2003. Optimal dispatching model for shenzhen
and management. Environ. Model. Softw. 24 (7), 786–797. water resources system. J. Hohai Univ. Nat. Sci. 31 (6), 616–620, (In chinese).

9126