IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

OPTIMAL DISTRIBUTION NETWORK 

RECONFIGURATION USING MULTI-OBJECTIVE 

CUCKOO SEARCH ALGORITHM 

AZRIN SAEDI1, MOHD SHAHRIN ABU HANIFAH1*, HILMI HELA LADIN2  

AND SITI HAJAR YUSOFF1 

1
Department of Electrical and Computer Engineering,  

2
Departmentt of Mechanical Engineering, 

International Islamic University Malaysia, 

Jalan Gombak, 53100 Kuala Lumpur, Malaysia. 

*
Corresponding author: shahrin@iium.edu.my 

(Received: 28th September 2021; Accepted: 16th January 2022; Published on-line: 4th July 2022) 

ABSTRACT:  In power system electricity delivery, the distribution system has the most 

electricity loss as the system has the highest R/X ratio and has a radial network at one time. 

Optimal reconfiguration of the distribution network is needed in order to reduce power 

losses. However, as it is also involved with multiple objectives and constraint problems 

such as switching frequency, voltage, and current limits, it is difficult to find the optimal 

solution. Hence, this paper proposes the Multi-objective Cuckoo Search (MOCS) 

algorithm to find the optimal reconfiguration of distribution networks by considering 

minimizing power losses and switch operations. Based on the simulation results on the 

IEEE-33 bus system, the performance of the MOCS-based scheme has been found to be 

significantly better than the single-objective algorithm thereby reducing approximately 

33% of the power losses.  

ABSTRAK Melalui sistem penghantaran jana kuasa elektrik, sistem pengagihan 

mempunyai pembaziran tenaga elektrik terbesar kerana sistem ini mempunyai nisbah R/X 

paling tinggi dan mempunyai satu rangkaian radial pada tiap-tiap satu masa. Konfigurasi 

semula rangkaian pengedaran yang optimum diperlukan bagi mengurangkan pembaziran 

tenaga. Walaubagaimanapun, oleh kerana ia melibatkan objektif dan kekangan masalah 

yang pelbagai seperti kadar peralihan, had voltan serta arus, adalah sukar bagi 

mendapatkan bacaan yang optimum. Oleh itu, kajian ini mencadangkan Carian Cuckoo 

Pelbagai Objektif (MOCS) bagi mencari konfigurasi semula yang optimum bagi sistem 

pengagihan tenaga dengan mengambil kira pengurangan pembaziran tenaga dan kadar 

peralihan. Berdasarkan keputusan simulasi pada sistem bas IEEE-33, pretasi MOCS telah 

menunjukkan peningkatan yang ketara berbanding algoritma objektif tunggal dan 

pengurangan sebanyak 33% tenaga.  

KEYWORDS:  distribution network reconfiguration (DNR); multi-objective Cuckoo 

search (MOCS) algorithm; power loss reduction; switch operations; 

Pareto optimal 

1. INTRODUCTION 

An efficient and high-reliability power system is crucial because global electricity 

demand has increased due to the rising population. Since the primary source of power 

around the world comes from non-renewable energy sources such as fossil fuels, coal, and 

natural gas, power loss in electricity delivery can be seen a contributor to global warming 

114



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

[1]. The distribution system has the most power loss compared with other systems as it has 

a high R/X ratio. This ratio means the reactance is much higher in the system [2]. System 

reconfiguration can be manipulated by two methods, power line restructuring and switch 

opening and closing. The tie switch (normally open) and sectionalizing switch (normally 

closed) are used to connect and disconnect power to the load entity to reduce the power loss 

in the distribution network. The procedure of changing the open/closed switch status is 

called distribution network reconfiguration (DNR). DNR is primarily performed to 

minimize losses and maximize the load balancing, system reliability, and voltage profile at 

the network level [3]. To manually change the status of the switches requires much time and 

cost and consumes energy.  

In the literature, two approaches have been used in optimizing this problem, which are 

using (i) a single-objective approach [4-8]  and (ii) a multi-objective approach [9-12]. For 

the single-objective approach, the author in [4] has focused on minimizing the power/energy 

losses and network loading index using a hybrid heuristic-genetic algorithm. While authors 

in [5,6] have compared several heuristic algorithms for minimizing power loss and 

enhancing voltage profile and applied them with to various scales of the distribution 

networks. At the same time, [5] has considered the integration of distributed generation 

(DG) in the network. Other researchers have investigated the optimal placement and sizing 

of unified power quality conditioners (UPFC) along with DNR for real power loss reduction 

[7]. However, the common drawback of these approaches is that the values of variables and 

parameters in each objective function vary depending on the case study and type of the 

network. Besides, it requires a weight factor in the objective function, which needs to be 

tuned to get the optimal solution.   

On the other hand, multi-objective optimization is a technique when two or more 

objectives are considered simultaneously to achieve the most desired outcome. This 

technique is used in many fields like economics, logistics, management, science, and 

engineering [13]. Unlike single-objective, this approach does not require a weight factor in 

the objective function. There are trade-offs among the objectives to achieve the outcome as 

the objectives might conflict with each other [9]. In DNR, research has been conducted 

using multi-objective approaches such as Multi-objective Evolutionary Algorithm [10] and 

Non-sorting genetic algorithm [11] for service restoration, and Bayesian learning-based 

evolutionary algorithm for absorption rate of wind power and voltage stability improvement 

[12]. However, less work has been reported on minimizing power losses and switching 

operations simultaneously using single-objective or multi-objective approaches. 

Hence, this paper proposes a solution for power loss reduction and switching 

minimization using the Multi-objective Cuckoo Search (MOCS) algorithm. MOCS is the 

extended version of the cuckoo search algorithm developed by Yang and Deb [14]. It has 

been tested against relevant test functions and then successfully applied to numerous 

problems [15-18]. This study uses MOCS to solve the DNR optimization problem focusing 

on power loss reduction and switching operations. The optimal model is developed in a 

MATLAB environment using the IEEE-33 bus test system. Then the result is compared to 

a single objective cuckoo search algorithm.  

2. PROBLEM FORMULATION

In this paper, the DNR problem is formulated as a multi-objective and multi-constrained

problem. The various objective functions and constraints considered in this work are 

explained as follows. 

115



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

2.1  Objective Functions 

The following equations express the objective functions aimed at minimizing the power 

loss and number of switch operations. 

a) Minimization of power losses:                        

min 𝑓1 = ∑ 𝑅𝑖
𝑃𝑖

2 + 𝑄𝑖
2

𝑉𝑖
2

𝑁𝑏𝑟

𝑖

                                                        (1) 

where Nbr is the total number of branches, Ri is the branch resistance i, Vi is the voltage at 

sending end node of ith branch, and Pi and Qi are the active and reactive power at the sending 

end node of ith branch.  

b) Minimization of switch operations: 

min 𝑓2 = ∑|𝑆𝑊𝐵𝑗 − 𝑆𝑊𝐴𝑗 |

𝑁𝑠

𝑗

                                                 (2) 

where Ns is the number of operated switches, SWAj and SWAj are the status of jth operated 

switch in the network before and after reconfiguration. 

2.2  Constraints 

The followings are the constraints that secure an optimal power flow calculation and 

preserve the network radial condition [6].  

a) Voltage Limit  

𝑉𝑖,𝑚𝑖𝑛 ≤ 𝑉𝑖 ≤ 𝑉𝑖,𝑚𝑎𝑥                (3) 

where 𝑉𝑖,𝑚𝑖𝑛 is 0.9 p.u and 𝑉𝑖,𝑚𝑎𝑥 is 1.1 p.u which is the voltage limit at end node of ith 
branch. 

b) Current Limit 

       𝐼𝑖 ≤ 𝐼𝑖,𝑚𝑎𝑥                    (4) 

where 𝐼𝑖 is the current at ith branch and 𝐼𝑖,𝑚𝑎𝑥  is the maximum current at ith branch. 

c) Radial Topology Constraint 

In any network, the number of main loops can be calculated using the following relation: 

           𝑁𝑛𝑜𝑑𝑒 − 𝑁𝑏𝑟𝑎𝑛𝑐ℎ = 1                        (5) 

where the configuration is radial, and the system has no isolated node. The number of 

nodes in the system is notated as 𝑁𝑛𝑜𝑑𝑒 . 

3.  METHODOLOGY 

3.1  Multi-objective Cuckoo Search (MOCS) Algorithm 

The Cuckoo Search algorithm is based on the aggressive way that cuckoo birds use to 

sustain the survival of their species. The bird would lay eggs in a host nest, and the survival 

of the eggs depends on the probability that the host bird discovers the eggs. The host bird 

would either abandon the nest or throw the eggs if they are discovered. Cuckoo birds 

develop the way to survive by mimicking the appearance of the host bird egg or the egg 

116



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

hatched earlier from the host bird egg. The pseudocode and the flowchart of MOCS is shown 

in Fig. 1 and Fig. 2 respectively. 

 

Initialize objective functions 𝑓1(𝒙), … , 𝑓𝐾 (𝒙)𝒙 = (𝑥1, … , 𝑥𝑑 )
𝑇  

Generate an initial population of 𝑛 host nests 𝒙𝑖  and each with 𝐾 eggs 
while (𝑡 < MaxGeneration) or (stop criterion) 
 Get a cuckoo (say 𝑖) randomly by Lévy flights 
 Evaluate and check if it is Pareto Optimal  

 Choose a nest among 𝑛 (say 𝑗) randomly 
 Evaluate 𝐾 solutions of nest 𝑗 
 if  new solutions of nest 𝑗 dominate those of nest 𝑖, 
  Replace nest 𝑖 by the new solution set of nest 𝑗 
 end 

 Abandon a fraction (𝑝𝑎 ) of worse nests 
 Keep the best solutions (or nest with non-dominated sets) 

 Sort and find the current Pareto optimal solutions  

end 

Postprocess results and visualization  

Fig. 1: Multi-objective Cuckoo Search Algorithm Pseudocode [14]. 

Start

Initialize objective functions

Generate initial population of n host nests and K eggs

Converge?

Get a cuckoo randomly by Levy flights

Evaluate and check if it is Pareto Optimal

Choose a nest among n randomly

Evaluate solutions

New solution dominated?

Replace chosen nest by new solution

Rand < Pa

Replace the worst solutions

Keep best solution

Sort and find current optimal solutions

Postprocess results and visualization

End

No

No

Yes

Yes

No

Yes

 

Fig. 2: Multi-objective Cuckoo Search algorithm flowchart. 

Cuckoo Search algorithm has three general rules:  

1) Each cuckoo bird lays one egg in only one random nest at a time;  

2) The best egg would survive for the next generation;  

3) The number of available host nests is the same; the probability of laid cuckoo egg to 

be discovered is between 0 and 1. The host bird would throw away the bird egg or 

abandon the nest, building a new nest [19]. 

117



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

In the multi-objective Cuckoo Search Algorithm, the first and third general rules are 

changed: 

1) Each cuckoo bird lays K eggs in only one random nest at a time; K refers to the 
number of objectives. 

2) The best egg would survive for the next generation; 

3) The number of available host nests is the same. The probability of laid cuckoo K 
eggs to be discovered is between 0 and 1. The host bird would build a new nest with 

K eggs in respect of the egg differences. Diversity would happen by random mixing.  

3.2  Proposed Algorithm and Pareto Optimal 

The algorithm aims to obtain the Pareto Optimal with respect to some switch changes 

and power loss with the highest voltage stability limit. This algorithm is improved for Pareto 

Optimal from the MOCS algorithm in [18]. The algorithm steps are: 

1. Data of the bus system (e.g. branch, bus and load number) are obtained. 

2. Bus voltage and power loss (𝑃𝑙𝑜𝑠𝑠) are calculated by running the load flow program. 

3. Voltage stability limit is calculated by the formula: 

𝑉𝑆 =
1

𝜆
, while λ is the load value. 

4. A set of the initially closed switches is defined as 𝑅𝑖 = ( (𝑅𝑖 )1
1 ⋯ (𝑅𝑖 )𝑛𝑒

1 ). 

5. Parameters for the algorithm are set such as nest dimension (𝑛𝑑 ), nests number (n), 
switch opened dimension (𝑛𝑒 ), the probability to be discovered, and step size (α), 
lower limit and upper limit search space, and maximum iteration number (N). 

6. Search space 𝑛 × 𝑛𝑑  is generated randomly. Each row represents a solution, while 
every element represents power loss for every connection 𝑃𝑐  to the load. 

𝑃𝑐 = (

(𝑃𝑐 )1
1 ⋯ (𝑃𝑐 )𝑛𝑑

1

⋮ ⋱ ⋮
(𝑃𝑐 )1

𝑛 ⋯ (𝑃𝑐 )𝑛𝑑
𝑛

) 

 

7. Search space 𝑛 × 𝑛𝑒  is generated. Each row represents a set of switches opened in 
the distribution network, while every element represents a switch closed for every 

radial network, 𝑅𝑐 .𝑅𝑐 = (

(𝑅𝑐 )1
1 ⋯ (𝑅𝑐 )𝑛𝑒

1

⋮ ⋱ ⋮
(𝑅𝑐 )1

𝑛 ⋯ (𝑅𝑐 )𝑛𝑒
𝑛

) 

8. Then, reactive power (𝑄𝑙𝑜𝑠𝑠)′ would be added to the bus and each 𝑃𝑐  row, new power 
loss (𝑃𝑙𝑜𝑠𝑠)′  and bus voltage is calculated.  

(𝑄𝑙𝑜𝑠𝑠)′ =  𝑄𝑙𝑜𝑎𝑑 −  𝑃𝑐  

9. Then, power loss reduction is calculated as: 

∆𝑃  =  𝑃𝑙𝑜𝑠𝑠 − (𝑃𝑙𝑜𝑠𝑠)′ 

10. Then, switch change number (𝑗) is calculated as this conditional loop statement:  

for every element 𝑅𝑖  ≠ 𝑅𝑐   

𝑗 is incremented by 2 

11. The optimal value for minimum power loss (𝑓1)  with the lowest switch change (𝑓2)  
is obtained. 

12. The bus voltage value is accepted if the bus voltage value is within the range of 
acceptable value. 

118



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

13. Levy flight is applied to obtain a new solution. 

14. Then, step 6 to 9 is repeated. 

15. An unaccepted solution is abandoned, and a new solution is generated 

16. Step 12 is repeated. 

17. Iteration value is increased if it does not reach the maximum number from step 10. 

18. Then, if the voltage stability limit is not the maximum value, the value of λ is 
increased, and step 3 is repeated. Otherwise, the algorithm is terminated. 

19. Then, all surviving solutions of the Cuckoo Search are plotted using the Pareto 
optimal. 

3.3   IEEE-33 Bus Test System 

The effectiveness of the MOCS algorithm has been studied on the standard IEEE-33 

bus model system [12] in a MATLAB environment using the MATPOWER package [20]. 

This test model is a power distribution system with 33 busses attached to the load points, 

and it is connected with 37 switches (sectionalize and tie switches), as shown in Fig. 3. Each 

load is identified by the number given, and in one mainline, the load number must be in 

sequence. Any tie switch would connect one point of the load with another point of the load 

of the different power lines. This tie switch would be closed to accommodate the breakdown 

of other sectionalizing switches to make sure every load is receiving power. The tie switch 

and sectionalizing switch can be alternately switched on and off to optimize power loss. 

This model system can be assumed to have constant base power, Sbase = 50 MVA, and base 

voltage, Vbase = 33 kV. The real power of the load is 3.715 MW, and the reactive power of 

the load is 2.3 MVAR. Minimum and maximum per-unit voltages are 0.95 p.u and 1.05 p.u, 

respectively. The initial power losses recorded in the model system is 208.46 kW. 

2         3       4       5       6       7         8     9      10     11  12       13     14     15     16     17     18  

26        27     28     29     30     31     32     33       

23        24     25          

19     20     21    22          

Substation
Load
Sectionalize switch
Tie switch

s1    s2     s3      s4     s5     s6      s7     s8     s9    s10   s11   s12    s13   s14   s15   s16   s17

 s26   s27    s28   s29   s30   s31   s32

 s23   s24

 s19   s20   s21

 s33

 s35

 s36

 s34

 s37

 

Fig. 3: The initial state of the IEEE-33 bus system [12]. 

4. RESULTS AND DISCUSSION 

In this paper, the MOCS algorithm is used to obtain the optimal DNR for minimizing 

the power losses and switch operation simultaneously. The result is analyzed using the 

Pareto optimal front, and the voltage profile for all solutions are presented. Then, it is 

compared to a single objective approach cuckoo search algorithm. Finally, the performance 

of MOCS is analyzed using the convergence rate in terms of generalized distance [14] and 

validated against the multi-objective genetic algorithm (MOGA) [21].  

119



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

4.1 Pareto Optimal Front 

The Pareto front generated by four non-dominated solutions after 50 iterations is 

presented in Fig. 4. The graph is aligned with Pareto optimal as it shows that the 

higher number of switch changes, the lower the power loss in the distribution 

network. The number of operated switches ranges between two and eight, contrary 

to the broadest possible range between zero and ten. At the same time, the power 

losses ranged from 138 kW to 156 kW. The detailed result for each Pareto point is 

tabulated in  

Table 1. From the table, all solutions show a significant reduction in power losses 

after obtaining optimal reconfiguration using MOCS. The lowest power loss is obtained by 

solution 1, which decreased from 208.46 kW to 138.93 kW, bringing approximately 33% 

reduction. It requires eight switches to be operated, which are number 7, 9, 14, 32, 33, 34, 

35, and 36. On the other hand, the lowest number of operated switches is recorded by 

solution 4, which only changed switches number 7 and 35. However, the power losses 

decreased to 155.80 kW, equivalent to a 25% reduction, which is slightly lower.  

 

Fig. 4. Pareto Optimal Multi-Objective Cuckoo Search Algorithm. 

 

Table 1: Simulation before and after reconfiguration using MOCS 

 
Before 

Reconfiguration 

After Reconfiguration 

Solution 1 Solution 2 Solution 3 Solution 4 

Tie Switch 33 34 35 36 37 7 9 14 32 37 7 9 14 36 37 7 11 34 36 37 7 33 34 36 37 

Power Loss 208.46 kW 138.93 kW 141.43 kW 143.80 kW 155.80 kW 

Power Loss 

Reduction 

- 33.36 % 32.15 % 31.02 % 25.26 % 

No. of 

Operated 

Switch 

- 8 6 4 2 

Minimum 

Voltage 

0.911 p.u 0.942 p.u 0.938 p.u 0.938 p.u 0.937 p.u 

  

120



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

4.2  Voltage Profile 

Fig. 5 presents the voltage profile of IEEE-33 bus system for each solution obtained by 

MOCS. The graphs show that the voltage profiles are significantly improved after the 

network is reconfigured across all solutions. Besides, the minimum voltages are also 

increased by approximately 3% compared to before reconfiguration. 

 

       

(i) Solution 1                    (ii) Solution 2 

       

(iii) Solution 3                   (iv) Solution 4 

Fig. 5: Voltage profile of the network before and after reconfiguration for each solution. 

4.3  Comparison with Single-objective Algorithm 

In order to see the effectiveness of the multi-objective technique, the above results are 

compared with the single-objective approach [22] demonstrated in Error! Reference 

source not found.. It is apparent that the power loss obtained by the single-objective is 

157.65 kW which is almost 1% lower power loss reduction compared to the solution 4 

recorded by MOCS in  

Table 1. Furthermore, in terms of switch operations, both single-objective and multi-

objective approaches were at the same level. Overall, MOCS shows better effectiveness 

compared to the single-objective algorithm. Besides, it provides multiple solutions to be 

chosen and does not require a weighting factor as a single-objective method. 

  

121



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

Table 2: Simulation Result of single objective Cuckoo Search Algorithm 

 Before Reconfiguration After Reconfiguration 

Tie Switch 33 34 35 36 37 7 34 35 36 37 

Power Loss 208.46 kW 157.65 kW 

Power Loss Reduction - 24.38 % 

No. of Operated Switch - 2 

Minimum Voltage 0.911 p.u. 0.930 p.u. 

4.4 Generational Distance (GD) Measurement 

Furthermore, in order to see the proposed MOCS performance, we also tested the same 

problems using a similar multi-objective technique, which is MOGA [21]. The performance 

is measured in terms of generational distance (GD)[14]. GD is designed to measure the sum 

of adjacent distances of solutions sets obtained by different algorithms, especially multi-

objective evolutionary algorithms. The comparison of the convergence rates between 

proposed MOCS and MOGA is plotted in Fig. 6. This figure shows that MOCS converged 

slightly faster than MOGA even though there were opposite patterns shown at early 

iterations, which could be neglectable. Nevertheless, overall, MOCS delivers better 

performance than MOGA. 

 

Fig. 6: Convergence comparison between MOCS and MOGA. 

5. CONCLUSION 

Power losses are critical in electrical power systems due to their impact on system 

reliability. Thus, distribution network reconfiguration (DNR) is introduced to minimize 

power losses and switch operations. In this study, a multi-objective Cuckoo Search (MOCS) 

algorithm for finding the optimal reconfiguration is presented. The simulation result is 

validated on the standard IEEE-33 bus test system in the MATLAB environment, and 

MATPOWER package is used for power flow calculation. The results show that MOCS 

obtained approximately 25% to 33% reduction results of power losses where two to eight 

switches were operated. Compared to the single-objective approach, MOCS recorded 

slightly better losses reduction and was on par in terms of switch operations. Besides, the 

convergence comparison with a similar multi-objective technique, MOGA, indicates the 

122



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

 

superiority of MOCS to obtain better optimal reconfiguration. Hence, the objectives of this 

paper to minimize the power losses and switch operations is achieved.  

Furthermore, the obtained Pareto optimal front shows the trade-offs between two 

objectives in solving the DNR problem. In this case, the system operator can benefit from 

the results of the MOCS application to decide the priority between power losses and 

numbers of switch operations based on actual circumstances. As a way forward, this 

research may be extended to the large-scale networks such as IEEE-69 and IEEE-129 bus 

system models with the integration of renewable energy sources. 

ACKNOWLEDGEMENT 

This work was supported and funded by International Islamic University Malaysia (IIUM) 

under IIUM-UMP-UiTM Sustainable Research Collaboration Grant 2020 (SRCG20-009-

0009). 

REFERENCES 

[1] Surana K, Jordaan SM. (2019) The climate mitigation opportunity behind global power 

transmission and distribution. Nature Climate Change, 9(9): 660-665. 

https://doi.org/10.1038/s41558-019-0544-3 

[2] Singh B, Chauhan S, Reddy CC. (2017) Power loss minimization in electrical power 

distribution networks by use of hybrid reconfiguration method. 3rd International Conference 

on Condition Assessment Techniques in Electrical Systems (CATCON), IEEE. pp. 349-354. 

https://doi.org/10.1109/CATCON.2017.8280243 

[3] Thakar S, Vijay AS, Doolla S. (2019) System reconfiguration in microgrids. Sustainable 

Energy, Grids and Networks, 17100191. https://doi.org/10.1016/j.segan.2019.100191 

[4] Jakus D, Čađenović R, Vasilj J, Sarajčev P. (2020) Optimal Reconfiguration of Distribution 

Networks Using Hybrid Heuristic-Genetic Algorithm. Energies, 13(7): 1544. 

https://doi.org/10.3390/en13071544 

[5] Nguyen HD, Valeev IM. (2019) Improvement methods for solving the distribution network 

reconfiguration problem. Energetika, 64(4): 174-185. 

https://doi.org/10.6001/energetika.v64i4.3892 

[6] Landero A, Koziel S, Abdel-Fattah MF. (2019) Distribution network reconfiguration using 

feasibility-preserving evolutionary optimization. Journal of Modern Power Systems and Clean 

Energy, 7(3): 589-598.  https://doi.org/10.1007/s40565-018-0480-7 

[7] Gholami K, Karimi S, Dehnavi E. (2019) Optimal unified power quality conditioner placement 

and sizing in distribution systems considering network reconfiguration. International Journal 

of Numerical Modelling: Electronic Networks, Devices and Fields, 32(1): e2467.  

https://doi.org/10.1002/jnm.2467 

[8] Samadaei E, Khosravi A, Sheikholeslami A. (2017) Optimal Allocation of Active Power Filter 

On real distribution network for improvement of power quality by use of BBO: A case study. 

IIUM Engineering Journal, 18(1): 85-99.  https://doi.org/10.31436/iiumej.v18i1.688 

[9] Gunantara N. (2018) A review of multi-objective optimization: Methods and its applications. 

Ai Q, editor. Cogent Engineering, 5(1): 1502242. 

https://doi.org/10.1080/23311916.2018.1502242 

[10] Sanches DS, London Junior JBA, Delbem ACB. (2014) Multi-Objective Evolutionary 

Algorithm for single and multiple fault service restoration in large-scale distribution systems. 

Electric Power Systems Research, 110: 144-153.  https://doi.org/10.1016/j.epsr.2014.01.017 

[11] Shahrin M, Aoki H. (2016) Application of Multi-Objective Optimization for Service 

Restoration Problem in Distribution Systems. IEEJ Transactions on Power and Energy, 136(3): 

275-283.  https://doi.org/10.1541/ieejpes.136.275 

123



IIUM Engineering Journal, Vol. 23, No. 2, 2022 Saedi et al. 
https://doi.org/10.31436/iiumej.v23i2.2190 

[12] Zhong T, Zhang H-T, Li Y, Liu L, Lu R. (2020) Bayesian Learning-Based Multi-Objective

Distribution Power Network Reconfiguration. IEEE Transactions on Smart Grid, 12(2): 1174-

1184.  https://doi.org/10.1109/TSG.2020.3027290

[13] Khan MR, Badran M, Toha SF, Zainal Abidin Z. (2021) Multi-Objective Optimization of

Snake Robot in Serpentine Locomotion. IIUM Engineering Journal, 22(2): 364-383.

https://doi.org/10.31436/iiumej.v22i2.1691

[14] Yang X-S, Deb S. (2013) Multiobjective cuckoo search for design optimization. Computers &

Operations Research, 40(6): 1616-1624.  https://doi.org/10.1016/j.cor.2011.09.026

[15] Zainal MI, Yasin ZM, Zakaria Z. (2021) Optimizing Voltage Profile and Loss Minimization

using Multi Objective Cuckoo Search Algorithm. 2021 IEEE 11th IEEE Symposium on

Computer Applications & Industrial Electronics (ISCAIE), IEEE. pp. 116-122.

https://doi.org/10.1109/ISCAIE51753.2021.9431835

[16] Yasin ZM, Aziz NFA, Salim NA, Wahab NA, Rahmat NA. (2018) Optimal Economic Load

Dispatch using Multiobjective Cuckoo Search Algorithm. Indonesian J. Electrical Engin.

Comp Sci, 12(1): 168.  https://doi.org/10.11591/ijeecs.v12.i1.pp168-174

[17] Zhou X, Liu Y, Li B, Li H. (2017) A multiobjective discrete cuckoo search algorithm for

community detection in dynamic networks. Soft Computing, 21(22): 6641-6652.

https://doi.org/10.1007/s00500-016-2213-z

[18] Rao NT, Sankar MM, Rao SP, Rao BS. (2021) Comparative study of Pareto optimal multi

objective cuckoo search algorithm and multi objective particle swarm optimization for power

loss minimization incorporating UPFC. Journal of Ambient Intelligence and Humanized

Computing, 12(1): 1069-1080.  https://doi.org/10.1007/s12652-020-02142-4

[19] Yang X-S, Deb S. (2009) Cuckoo Search via Lévy flights. 2009 World Congress on Nature &

Biologically Inspired Computing (NaBIC), IEEE. pp. 210-214.

https://doi.org/10.1109/NABIC.2009.5393690

[20] Zimmerman RD, Murillo-Sanchez CE, Thomas RJ. (2011) MATPOWER: Steady-State

Operations, Planning, and Analysis Tools for Power Systems Research and Education. IEEE

Trans Power Sys, 26(1): 12-19.  https://doi.org/10.1109/TPWRS.2010.2051168

[21] Bevilacqua V, Pacelli V, Saladino S. (2011) A Novel Multi Objective Genetic Algorithm for

the Portfolio Optimization. Lecture Notes in Computer Science (Including Subseries Lecture

Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 186-193.

https://doi.org/10.1007/978-3-642-24728-6_25

[22] Fuad NAA, Azlin AAN, Zahari MFD, Hanifah MSA. (2018) Power Distribution Loss

Reduction Using Cuckoo Search. 2018 7th International Conference on Computer and

Communication Engineering (ICCCE), IEEE. pp. 333-337.

https://doi.org/10.1109/ICCCE.2018.8539252

124