Professional Documents
Culture Documents
Benavides Et Al 2013 Iterated Algorithm Appl Electrical Network
Benavides Et Al 2013 Iterated Algorithm Appl Electrical Network
a r t i c l e i n f o a b s t r a c t
Available online 24 May 2012 We present a metaheuristic approach which combines constructive heuristics and local searches based
Keywords: on sampling with path relinking. Its effectiveness is demonstrated by an application to the problem of
Metaheuristics allocating switches in electrical distribution networks to improve their reliability. Our approach also
Sample algorithms treats the service restoration problem, which has to be solved as a subproblem, to evaluate the
Path relinking reliability benefit of a given switch allocation proposal. Comparisons with other metaheuristics and
Power systems reliability with a branch-and-bound procedure evaluate its performance.
Switch allocation & 2012 Published by Elsevier Ltd.
n
We use an undirected graph G ¼(N,E) to model a distribution
Corresponding author. Tel.: þ55 5133086818.
E-mail addresses: ajbenavides@inf.ufrgs.br (A.J. Benavides),
network. The set of nodes N represents substations and consumer
marcus.ritt@inf.ufrgs.br (M. Ritt), buriol@inf.ufrgs.br (L.S. Buriol), load points, and the set of edges E represents the feeder lines.
paulo.morelato@fct.unesp.br (P.M. Franc- a). A Boolean value Be indicates the presence of a switch on a line
1 2 3 1 2 3
a
S(f)
8 13 8 13
4 10 14 4 b 10 14
c
S(9)
9 9
5 11 12 5 11 12
d
6 7 16 15 6 7 16 15
Fig. 1. Three-feeder example of a distribution network by Civanlar et al. [4]. (a) Normal operating conditions and (b) sectors affected by a failure.
e A E. We represent a solution for the switch allocation problem by area after the isolation of a failure. A solution has to satisfy electrical
a set EB ¼ feA E9Be ¼ 1g of lines that are selected to install new constraints. The electrical constraints considered in this paper are
switches. The sector SðnÞ of a node n A N is the largest connected the radiality of the reconfigured network, the capacities of the lines
subgraph which contains n and is connected only with basic and substations, and the maximum allowed voltage drop.
circuit feeder lines (normally closed) that have no switch installed Since we are interested in solving the service restoration
(Be ¼0). Ellipses in Fig. 1b represent sectors. The sector of a line problem as a subproblem of the switch allocation problem, we
e A E is defined as the union of e with the sectors corresponding do not consider secondary objective functions, such as minimiz-
to the two incident nodes. The frontier F ðG0 Þ of a subgraph ing the number of switching operations. For the same reason, we
(or sector) G0 is the set of edges e A E which are incident to exactly do not consider additional constraints such as the priority of
one node in G0 . customers, switching times, or restoration cost.
For a single substation, the service restoration problem can be
2.2. Network reliability estimation modeled as finding a spanning tree minimizing the non-supplied
energy subject to electrical constraints [6]. Different from the
In this paper, we use the expected energy non-supplied (EENS) minimum weight spanning tree, this version is NP-hard [7,8].
[5] to measure the network reliability. It is defined as Our service restoration algorithm calculates the affected con-
X X sumers after a failure isolation. It expands the attended area
EENS ¼ lf rf P n ðMW h=yearÞ,
starting from the substations sector by sector, when this is
f A Enc n A Nf
possible without violating the electrical constraints. The electrical
where Enc is the set of lines that can fail (normally closed), Nf is simulation is computationally expensive, but important to obtain
the set of nodes affected by a failure f, lf is the average failure a realistic approximation of the attended area.
rate, rf is the average outage time, and Pn is the average energy A detailed description of the reliability estimation and the
consumption of node n. service restoration algorithm can be found in [9]. For a survey of
Our reliability estimation algorithm simulates a failure f in approaches to service restoration we refer the reader to Ćurčić
each sector S of the network, and opens the switches of the et al. [10] and Sudhakar and Srinivas [11].
corresponding frontier F ðSðf ÞÞ to isolate the failure. Then, it
determines the set Nf of consumers affected by the failure with
a service restoration algorithm, and calculates their partial EENS. 2.4. The switch allocation problem
Evaluating sector by sector, the partial reliability of every feeder
line of a sector is evaluated at once, reducing the number of The number and position of the switches in the network
service restoration problems that have to be solved. The same influences the amount of non-supplied energy in case of con-
process is applied to the remaining lines that are not part of any tingencies. The switch allocation problem consists in selecting a set
sector. of feeder lines to install a given number of new switches in a
distribution network. The objective is to maximize the reliability,
2.3. The service restoration problem and is subject to electrical and topological constraints.
The switch allocation problem is NP-hard, since its solution
Network reconfiguration is the process of opening and closing requires to solve the NP-hard service restoration problem. For this
some switches to change the network topology. When a power reason the literature concentrates on heuristic solution methods.
failure is detected, the network is reconfigured to isolate the Exact methods for the switch allocation problem based on
failure and to restore energy by alternate lines. integer programming have been proposed by Zambon et al. [12],
Consider a failure in line f of Fig. 1b. Without switches, the Bupasiri et al. [13], Soudi and Tomsovic [14,15], Sohn et al. [16].
whole tree under substation 2 would be unattended. When the These methods apply only to small-sized distribution networks.
frontier switches F ðSðf ÞÞ ¼ fa,b,cg are opened, the failure is iso- Among the heuristics proposed are simulated annealing
lated in sector Sðf Þ. Consequently, sector Sð9Þ has neither failure [17], immune algorithms [18,19], genetic algorithms [20–22],
nor power supply. The service on sector Sð9Þ is restored when the divide-and-conquer [23], reactive tabu search [24], evolutionary
switch on line d is closed. algorithms [25], particle swarm optimization [26], and ant colony
The service restoration problem consists in choosing which optimization [5]. Some of these methods address the more
switches must be opened or closed to minimize the unattended general problem of network reinforcement planning, allowing to
26 A.J. Benavides et al. / Computers & Operations Research 40 (2013) 24–32
install besides the sectionalizing switches other devices, e.g. 3.2. Local search algorithms
protective devices such as reclosers or fuses, or capacitor banks.
Comparing the efficiency of these methods directly is not A local search algorithm iteratively replaces the current solu-
possible, since the authors use different test instances, which tion with a better neighbour. The entire neighbourhood for our
are often not completely specified or from private real networks. local search strategies is defined by the relocation of one switch at
Further, reliability is measured differently, for example by SAIDI a time from its current position e A EB to a feeder line e0 A E\EB .
[20], SAIFI [21], EENS [23,5], or ECOST [17,26]. Some authors use First improvement local search explores the neighbourhood until
combined costs for non-supplied energy and installation and a better neighbour is found. This neighbour becomes the current
maintenance [17–19,22–24,26,5], or multiple objectives, e.g., the solution and the search continues with the next iteration. The
number of installed switches versus the non-supplied energy number of evaluated neighbours varies with the iterations. Best
[25], or the installation costs versus the EENS [5]. improvement local search explores all the neighbourhood to
Many of these methods considerably simplify the underlying select the best neighbour for the next iteration. Sample local
service restoration problem by assuming that any redundant line search (Fig. 3) reduces the size of the explored neighbourhood
can restore energy supply and that the affected area or the overall and the number of reliability estimations. It samples b percent of
cost are known or easy to compute. the lines with switches ða A EB Þ and b percent of lines without
switches ðbA E\EB Þ to relocate one switch. If the algorithm finds a
better solution in the sample, it becomes the current solution for
the next iteration. Sample and best improvement evaluate a
3. Construction and local search
constant number of neighbours in each iteration, but sample
search reduces the neighbourhood size.
This section describes four construction algorithms (random,
First and best improvement stop when there are no better
sample, greedy and semi-greedy) and three local search strategies
solutions in the neighbourhood. Sample neighbourhood explora-
(sample search, first and best improvement) for the switch
tion is not exhaustive and does not guarantee to find a local
allocation problem.
minimum. Thus, the stop criterion may be a maximum number of
iterations or a number of iterations without improvement. To
3.1. Construction algorithms guarantee that a local minimum is reached, a best improvement
local search can be performed after the sample search.
Random construction selects k switches randomly and evalu- When installing k switches, the number of reliability estima-
2
ates the resulting solution. Greedy, semi-greedy and sample tions of each iteration is Oðk9E9b Þ for the sample local search and
constructions build a feasible solution by adding one element at Oðk9E9Þ for the other two strategies.
each iteration. Greedy construction evaluates all the elements to
select the best each time. Semi-greedy construction (Fig. 2a)
evaluates every possible element to build a restricted candidate 4. Iterated sample construction with path relinking (ISCPR)
list which contains a fraction of a elements with the highest
reliability. Finally, one element is selected randomly from the ISCPR is a variant of iterated local search (ILS) [27]. At each
restricted candidate list. (Setting a ¼ 0 selects the best element, iteration, ISCPR perturbs the last solution with a partial random
while a ¼ 1 selects randomly between all the elements.) Sample destruction followed by a sample construction, and uses path
construction (Fig. 2b) first selects randomly a sample of b relinking as a guided local search. Path relinking [28] explores the
elements. Then, it evaluates the sample candidate list to choose neighbourhood between two solutions. It repeatedly relocates
the best. (A value of b ¼ 0% corresponds to a random construc- one switch of the initial solution, to a position defined by the
tion, while b ¼ 100% corresponds to a greedy construction). guiding solution. It can be seen as a local search where the
The most expensive operation is the reliability estimation. neighbourhood is limited to the differences between initial and
When installing k switches, the number of reliability estimations guiding solutions. By sampling and searching only in a limited
is Oðk9E9Þ for the greedy and semi-greedy constructions, Oðk9E9bÞ neighbourhood, ISCPR reduces the number of reliability estima-
for sample construction, and O(1) for random construction. tions during construction and local search phases.
Table 1
Instances.
BR 1 32 39 3.70 12.66
BU 1 32 39 3.70 12.66
R3 1 32 37 4.55 12.66
B4 3 38 72 24.58 11.00
R4 11 83 96 35.20 12.66
AR 5 80 109 28.60 20.00
AU 5 80 109 28.60 20.00
R5 8 135 156 19.96 12.66
R6 3 201 216 32.44 33.60
R7 7 873 900 148.99 126.60
Table 1 presents details of the problem instances that were used To compare construction and local search methods, we tested
in our tests. Instances AU and AR are from Augugliaro et al. [7], all combinations C–L of construction algorithms and local search
and BU and BR are from Baran and Wu [6]. The total power strategies. The construction algorithm C can be greedy (Gr), semi-
demand was divided between the consumers load points. greedy (SGr), sample (Spl), or random (Rnd). The local search L
28 A.J. Benavides et al. / Computers & Operations Research 40 (2013) 24–32
Table 2
Parameters and values used in the experiments.
70
19 Construction
Rnd
Local search
17 BI 50
FI
EENS (MWh/year)
Spl
16
40
15
30
14
20
13
10
12
0
0 2 4 6 8 10
Time (seconds)
240
6 Construction
Rnd
Relative deviation from best known solution (1.793 MWh/year)
SGr
5.5 Spl
Gr 200
5 Local search
BI
FI 160
4.5 Spl
EENS (MWh/year)
4 120
3.5
80
3
2.5 40
2
0
0 20 40 60 80 100
Time (seconds)
Fig. 5. Average performance for construction and local search combinations. (a) Instance B4 and (b) instance R6.
A.J. Benavides et al. / Computers & Operations Research 40 (2013) 24–32 29
can be first improvement (FI), best improvement (BI), sample local minimum increases the runtime, but it remains about half of
(SLS), or sample followed by best improvement (SBI). For exam- the runtime of BI.
ple, Spl–SBI is the combination of a sample construction and a Rnd-FI is the most expensive combination. The cheapest
sample local search followed by a best improvement local search. combination is Spl–Spl. The fastest combination that finds the
We repeated each experiment 1000 times for B4, and 100 times best known solution in our experiments is Spl–SBI.
for R6 (except Gr–BI which is deterministic) allocating 20
switches. Fig. 5 and Table 3 report average results.
Table 3 shows the average relative deviation from the best 5.4. Comparison of metaheuristics
known solution after the construction (c) and after the local
search (l), the average runtime after construction (tc) and local In this section, we compare ISCPR with a Greedy Randomized
search (tl), and the percentage of tests that reach the best known Adaptive Search Procedure (GRASP) [3]. GRASP is a metaheuristic
solution (lBK), or the relative deviation of the best found solution which repeatedly constructs a solution and applies a local search.
in parentheses when the best known solution is not found. The Both the semi-greedy and the sample construction algorithms
best solution found in our tests for 20 switches has an EENS of present the greedy, randomized and adaptive characteristics of
11.262 MW h/year for B4 and 1.793 MW h/year for R6. GRASP as described below. The semi-greedy construction restricts
Fig. 5 shows the average EENS and the relative deviation (in greedily the candidate list and selects randomly one of them,
percent) from the best known solution over time. Four points while the sample construction restricts randomly the candidate
show the average result of the construction algorithms and curves list and selects greedily one of them. Both constructions are
depict the average performance of the local search methods. adaptive when selecting a new element because they take into
In the following we refer to values for instance B4 and give the account the benefits of the previously selected elements. Four
values for R6 in parentheses. Initial solutions created by the semi- variants of GRASP were implemented combining the construction
greedy algorithm are better than random solutions by 2.2 and local search algorithms tested before (SGr–BI, SGr–SBI, Spl–BI,
(1.2) MW h/year, and require 0.5 (4.7) s while random solutions Spl–SBI). We also tested variants of GRASP with path relinking to
require less than 4 ms. Greedy construction generates always the the current best solution after the local search (GRASP-PR), since
best initial solution at almost the same computational effort than often this can improve GRASP [31].
semi-greedy construction, but a following local search does not An iterated semi-greedy construction with path relinking, ICPR
find the best known solution. Solutions created by the sample (SGr-PR) is also compared to ISCPR (Spl-PR). Fig. 6 and Table 4
algorithm are better than random solutions by 5.0 (3.4) MW present results for the installation of 20 switches. Experiments
h/year. Sample construction creates better solutions than the were repeated 100 times, except for instance R7 with 10 repeti-
semi-greedy algorithm in about 10% of the corresponding time, tions). The stop criterion is a fixed time tmax given in Table 4, and
having the best cost/benefit among all construction algorithms. 10 min for results in Fig. 6.
The average final solutions after local search shown in Table 3 Fig. 6 presents a time-to-target plot [32,33] which shows the
are less than 3.5 (4.6)% from the best known solution, except after percentage of tests that reach a given target value within a given
sample local search with 6.9 (12.8)%. All search strategies find the execution time. It uses instance B4. The target is the best known
best known solution at least once, except sample search, which solution for 20 switches.
yields solutions 0.4 (0.9)% above the best known value. The search Comparing the average runtime of each heuristic to standard
strategies differ in their performance over time. The curves in GRASP (SGr-BI), the sample search GRASP (SGr-SBI) is faster by a
Fig. 5 show a quick initial progress for FI, but BI becomes better factor of 0.75, the sample construction GRASP (Spl–BI) is slower by a
later. FI finds better neighbours in early iterations, but spends factor of 2.4, GRASP with both sample algorithms (Spl–SBI) is slower
most of its runtime in later iterations with small improvements. by a factor of 1.8, and GRASP-PR (SGR-BI-PR) is slower by a factor of
The runtime of sample search is very small, due to the restricted 1.4. The other GRASP-PR variants that we tested took longer than
neighbourhood. Performing BI after sample search (SBI) to reach a their GRASP counterparts to obtain the same percentage of target
Table 3
Comparison of construction and local search algorithms.
Gr BI 3.95 0.826 3.95 (3.95) 2.51 7.37 6.450 0.06 (0.06) 113.54
SGr FI 49.53 0.444 2.18 17.5 29.56 164.41 4.699 2.99 10.0 808.58
BI 2.24 21.3 18.19 3.51 5.0 309.60
SLS 6.18 0.5 1.12 11.34 (1.06) 15.33
SBI 2.64 15.1 9.01 2.78 1.0 152.72
Rnd FI 69.20 0.001 2.40 8.6 44.42 229.47 0.003 4.56 1.0 1182.59
BI 2.36 14.5 22.11 3.79 4.0 328.93
SLS 6.22 0.2 0.74 12.76 (2.14) 11.04
SBI 2.76 13.5 8.53 3.66 4.0 137.41
Spl FI 25.04 0.056 3.37 5.3 18.98 43.01 0.584 2.76 15.0 351.71
BI 3.33 7.8 13.12 2.57 7.0 223.36
SLS 6.87 (0.41) 0.59 11.08 (0.91) 8.65
SBI 3.45 5.8 8.38 2.57 2.0 133.42
c, l: average percentage over best known solution of construction and local search, respectively.
lBK: percentage of tests that reach the best known solution (or relative deviation in parentheses if not reached).
tc, tl: average runtime in seconds of construction and local search, respectively.
30 A.J. Benavides et al. / Computers & Operations Research 40 (2013) 24–32
0.9
0.8
0.7
0.6
Probability
0.5
0.4
0.3
ISCPR (Spl-PR)
0.2 ICPR (SGr-PR)
GRASP (SGr-SBI)
GRASP (SGr-BI)
0.1 GRASP-PR (SGr-BI-PR)
GRASP (Spl-SBI)
GRASP (Spl-BI)
0
0 100 200 300 400 500 600
Time (in seconds) to target (11.262 MWh/year)
Table 4
Comparison of ISCPR with GRASP.
solutions; they are not shown in Fig. 6 for this reason. This indicates Table 4 compares ISCPR and standard GRASP (SGr-BI) for other
that the effectiveness of path relinking is reduced when applied to instances. It shows the best known solution value (BK), the time
local minima after a local search for this problem. limit (tmax), the number of iterations (Itmax) reached within the
ICPR (SGr-PR) performs slightly faster than the best GRASP time limit, the average relative deviation from the best known
combination. This results from the faster semi-greedy reconstruc- solution (S), the percentage of tests that reach the best known
tion that rebuilds only half of the switches, and the reduced solution (SBK), or the relative deviation of the best found solution
neighbourhood exploration of path relinking. ISCPR always in parentheses when the best known solution is not found, the
reaches the best known solution in less than 2 min, about 25 s average runtime (t), and the average iterations (It) needed to
in average. ISCPR is more than two times faster than all other reach a best solution.
tested combinations. This is a consequence of the fast execution ISCPR presents better average results than GRASP and reaches
of the sample reconstruction and of the guided search of path the best known solutions more often. Within the time limit, ISCPR
relinking over the reduced neighbourhood. iterates at least 10 times more than GRASP. Consequently, ISCPR
Sample local search reduces the search time of GRASP. Sample finds its best solution faster than GRASP. The better performance
construction also reduces the runtime, but it generates biased of ISCPR is more evident in large instances such as R7. It finds the
initial solutions and GRASP needs more iterations to compensate best solution in 130 min and 244 iterations in average for R7,
this. On the other hand, sample construction combined with path while GRASP completes only one iteration in an average of
relinking (ISCPR) reduces the runtime without loosing quality, 172 min. With the exception of instances BR and R3, ISCPR finds
and performs faster than semi-greedy construction combined better solutions faster and with higher probability. In particular,
with path relinking and than GRASP variants. for large instances with more than 80 lines, it finds the best
A.J. Benavides et al. / Computers & Operations Research 40 (2013) 24–32 31
known value in about half the execution time with a 13 times The use of path relinking as guided local search can improve the
higher average probability. solutions efficiently.
These components together form an efficient heuristic, ISCPR,
that is capable of reducing the number of reliability estimations,
5.5. Quality of the solutions and is able to perform better than GRASP for this problem.
improvement in distribution systems. In: Transmission and distribution [28] Glover F. Genetic algorithms and scatter search: unsuspected potentials.
conference and exposition: Latin America, 2004. IEEE/PES; 2004. p. 51–6. Statistics and Computing 1994;4(2):131–40.
[23] Carvalho PMS, Ferreira LAFM, da Silva AJC. A decomposition approach to [29] Kavasseri R, Ababei C. REDS: REpository of Distribution Systems. URL
optimal remote controlled switch allocation in distribution systems. IEEE /http://venus.ece.ndsu.nodak.edu/kavasseri/reds.htmlS; 2009.
Transactions on Power Delivery 2005;20(2):1031–6. [30] Benavides AJ, Machado M, Ritt M, Buriol LS. Problem instances for the switch
[24] da Silva LGW, Pereira RAF, Abbad JR, Mantovani JRS. Optimised placement of allocation problem. URL /inf.ufrgs.br/algopt/repos/power-distribution-in
control and protective devices in electric distribution systems through reactive stances.tgzS; 2011.
tabu search algorithm. Electric Power Systems Research 2008;78(3):372–81. [31] Resende M, Ribeiro C. GRASP with path-relinking: recent advances and
[25] Villasanti A, Baran B, Gardel P. Multiobjective allocation of remotely con- applications. In: Ibaraki T, Nonobe K, Yagiura M., editors. Metaheuristics:
trolled switches in an electric distribution power system. In: Transmission progress as real problem solvers. Operations research/computer science inter-
and distribution conference and exposition: Latin America, 2008. IEEE/PES;
faces series, vol. 32. Springer US; 2005. p. 29–63, ISBN 978-0-387-25382-4.
2008. p. 1–7.
[32] Aiex R, Resende M, Ribeiro C. TTT plots: a perl program to create time-to-
[26] Moradi A, Fotuhi-Firuzabad M. Optimal switch placement in distribution
target plots. Optics Letters 2007;1:355–66.
systems using trinary particle swarm optimization algorithm. IEEE Transac-
[33] Hoos H, Stützle T. On the empirical evaluation of Las Vegas algorithms—position
tions on Power Delivery 2008;23(1):271–9.
paper. Technical Report, Computer Science Department, University of British
[27] Stützle T. Applying iterated local search to the permutation flow shop
problem. Technical Report AIDA-98-04. Darmstadt, Germany: University of Columbia; 1998.
Technology, Department of Computer Science, Intellectics Group; 1998.