8/10/2019 06019279

1/8

1

Abstract--In recent years, a substantial amount of photovoltaic

(PV) generations have been installed in Japanese power systems.

However, the power output from the PV is random andintermittent in nature. Therefore, the PV generation poses many

challenges to system operation. To evaluate impacts of the

behavior of PV on power supply reliability, we have developed

power supply reliability evaluation model considering a large

integration of PV generation into power system. As a result, the

power supply reliability is getting worse as the PV penetration

increases. To mitigate this issue, we have proposed that pumped

storage hydro power plant (PSHPP) is used to improve the

reliability. However, its operation may increase the power system

operational cost. In this paper a new method for scheduling the

effective operating pattern for PSHPP that makes it possible to

improve both reliability and economy is presented.

Index Terms--Genetic Algorithm, Monte Carlo Simulation,Optimal Scheduling, Photovoltaic Generation, Power Supply

Reliability, Pumped Storage Hydro Power Plant, Surplus Power

Problem, Tabu Search.

I. INTRODUCTION

HE role of renewable energy resources is remarkable in

Low Carbon Society. In recent years, a large amount of

photovoltaic (PV) and wind power generations have been

installed in power systems around the world. However, Japan

is not an appropriate site to install a large amount of wind

power generations. Therefore, the Japanese government aims

at installing a large amount of PV in power system.

Since the power output from the PV is random andintermittent in nature, the PV generation poses many

challenges to the power system operation. When the load

demand is very small and at the same time the PV generates a

large amount of power, the balance between power supply and

demand cannot be maintained. This problem is called surplus

power problem". In addition to the surplus power problem, it

Ryota Aihara is with the Department of Electrical Engineering, The

University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 113-8656, Japan

(email: aihara@syl.t.u-tokyo.ac.jp).

Akihiko Yokoyama is with the Department of Advanced Energy, The

University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8561, Japan.

Fumitoshi Nomiyama and Narifumi Kosugi are with the Power System

Engineering Group, Research Laboratory, Kyushu Electric Power Company,Inc., Fukuoka, 815-0032, Japan.

is also necessary to consider the intermittent and uncertain

output characteristic of PV.

Therefore, various kinds of counter measures have been

considered when a substantial amount of PV generation is

installed into a power system. Though Battery Energy Storage

System (BESS) could be considered as a solution of these

issues, their high cost is preventing them from being

considered as a solution in most situations. This paper

proposes the effective use of Pumped Storage Hydro Power

Plant (PSHPP). PSHPP is installed originally for load leveling

within one day. The energy is stored in the form of water

pumped up from a lower reservoir to an upper reservoir. The

PSHPP is usually operated as a generator in the daytime and a

pump in the nighttime. A new operation scheduling of PSHPP

is proposed to solve this problem in this paper. It will be

operated with the proposed pattern solving the surplus powerproblem caused by the PV generation. The authors have

already proposed how the PSHPP operation pattern is changed

effectively to improve the power system reliability in case of a

large integration of PV into the power system [1]. However,

the scheduling method for obtaining the optimal PSHPP

operation pattern that makes it possible to improve both

reliability and economy has not been developed yet. In this

paper, we propose a new method for planning PSHPP

operation pattern considering the surplus power problem,

power supply reliability and fuel cost reduction, which is

represented by pareto optimal solutions. The power supply

reliability and the fuel cost are estimated for each PSHPP

operation plan by using Monte Carlo simulation.

II. POWER SUPPLY RELIABILITY EVALUATION MODEL

A. Pumped Storage Hydro Power Plant

In PSHPP, the energy is stored in upper reservoir in the

form of water. Therefore, it should be operated within the

capacity of the upper reservoir. In this paper, the amount of

the stored water in the upper reservoir is considered as energy.

It can be operated 6 to 8 hours continuously when the

operation starts from the full capacity of the upper reservoir.

Therefore, the capacity of the upper reservoir is assumed to be

7-hour operation of the rated power at generator mode.

Optimal Operation Scheduling of Pumped

Storage Hydro Power Plant in Power System

with a Large Penetration of Photovoltaic

Generation Using Genetic AlgorithmRyota Aihara, Akihiko Yokoyama,Member, IEEE, Fumitoshi Nomiyama, and Narifumi Kosugi

T

Paper accepted for presentation at the 2011 IEEE Trondheim PowerTech

978-1-4244-8417-1/11/$26.00 2011

8/10/2019 06019279

2/8

2

In general, there are two types of PSHPP, fixed speed one

and adjustable speed one. The fixed speed PSHPP cannot

adjust the input power when it operates as a pump. The

adjustable speed one can adjust the input power. It is coming

into the limelight because of its adjustable characteristics in

the pump mode [2]. Therefore, the PSHPP is assumed to be anadjustable speed PSHPP in this paper.

Fig. 1. Flowchart of Power Supply Reliability Calculation

B. Reliability Evaluation Model

The proposed evaluation method of the power supply

reliability is summarized as a flowchart in Fig.1. In this paper,

it is assumed that the operation schedule of PSHPP is planned

for a week. The simulation period is a week and its sampling

time is 1 hour. First, based on the available data, the

generation schedule for 7 days is determined. The available

data consist of load demand given by power system model,

theoretical PV generation output and PSHPP operation pattern.

The generation schedule is determined by solving

optimization problem, which minimizes the fuel cost by using

Dynamic Programming (DP). Next, Monte Carlo simulation is

carried out considering uncertainties of load demand, PV

output and generator failure. During the period of the

simulation, when the power supply and demand imbalance

occurs, the thermal generator outputs are adjusted and PSHPP

is operated to avoid the imbalance. If it cannot be avoided,

this period is regarded as supply interruption. We assume that

both supply shortage and surplus power supply are defined asthe supply interruption.

1) Generation Scheduling

As mentioned before, generation scheduling is found by

solving an optimization problem in which the fuel cost is

minimized. In this paper, it is assumed that nuclear and hydro

power plants are always operated at the rated power output.

As a result, only the fuel cost of thermal plants is considered

here. The fuel cost of a thermal plant is given in (1).

iiiiii cPbPaF 2 (1)

where ai

, bi

and ci

are fuel cost coefficients of generator i.Pi

ispower output of generator i.

The optimization problem for one-week period is then

formulated as follows:

Minimize:

})1()({ 1

168

1 1

ii

tt

N

i

i

t

i

t

ii

t SuuPFu

(2)

Subject to:

tHN

N

i

i

tt PVPPPL 1

(3)

ii

t

i PPP maxmin (4)

where is the total fuel cost of thermal power plants for one

week. N is the number of thermal power plants. Pti is power

output of generator iat time t. utiis state variable of generator

i at time t (1: committed 0: stopped). Si

is startup cost ofgenerator i.Ltis load demand at time t.PNis power output of

nuclear power plant.PHis power output of hydro power plant.

PVt is theoretical power output of PV at time t. Pimin is

minimum output of generator i. Pimax is maximum output of

generator i.

To solve the optimization problem, Dynamic Programming

(DP) [3] is used in this research. The merit order method is

used in the unit commitment. The merit order is determined by

using the fuel cost of generator at the rated output. The

combination of committed generators is searched sequentially

in ascending order of Cidescribed by (5).

max

max )(i

iii

P

PFC (5)

For simplification, the shutdown cost is not considered

here. The PV output is assumed to be theoretical in generation

scheduling because power system operator cannot predict its

output precisely at this time. The theoretical power output of

PV is shown in Fig. 2.

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

PVOu

tput[%]

Time Fig. 2. Theoretical PV Output

2) Monte Carlo Simulation

In this paper, we consider three types of probabilistic

uncertainty concerning load demand, power output of PV and

generator failure.

The load demand model with a randomly fluctuating

component is derived from (6). Ldt is load demand at time t

considering probabilistic fluctuation. Lgt is load demand at

time tgiven by the power system model and Fis the random

8/10/2019 06019279

3/8

3

number based on the normal distribution where the average is

1 and the standard deviation is 3%.

FLL gtdt (6)

The power output of PV generation is also modeled with a

randomly fluctuating component.

FPVPV gtdt (7)

where PVdt is PV output at time t considering probabilistic

fluctuation,PVgtis PV output at time tgiven by the theoretical

output and F is the random number based on the normal

distribution considering weather changes shown in Table 1.

TABLEI

RANDOM CHARACTERISTIC OF WEATHER CHANGESWeather Average Standard Deviation [%]

Sunny 1 3

Cloudy 0.5 20

Rainy 0.1 0.03

The weather change is simply modeled at the probability of

1/3 at time t. It is assumed that the weather is independent

between before and after time t. The theoretical PV outputs

for three kinds of weathers are shown in Fig. 3.

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

PVOutput[%]

Time

Sunny

Cloudy

Rainy

Fig. 3. PV Output for Three Kinds of Weather

The generator failure occurs during Monte Carlo

simulation at the failure rate described by (8). In (8), MTBFi

is the mean time between failure of generator iand MTTRiis

the mean time to repair of generatori

iiMTTR

1,MTBF

1 ii (8)

3) Generator Re-dispatch

During Monte Carlo simulation, if the generator failure

occurs at time t, the other generators are re-dispatched

immediately. Figure 4 shows how the generators are re-

dispatched when generator G3 is tripped at time t, for example.

First, Generator G1, G2 and G4 are re-dispatched immediately.

If imbalance of power supply and demand cannot be avoided

by the re-dispatching, PSHPP is operated immediately to

avoid it as an emergency control. In the next time t+1, the

remaining generators are committed to compensate for the

unavailable generator. The tripped generator is restored at the

restoration rateafter time t+1.

Time t

G1

LoadDemand

G2

G3

G4

G1

LoadDemand

G2

G4

Supply Demand

MaximumOutput

Supply

Shortage

G1

LoadDemand

G2

G4

MaximumOutput

Emergency

Control

PSHPP

F ir st Step Second S tep

Generators Re-dispatch

G1

LoadDemand

G2

G5

G4

Time t+1

Supply Demand

Generator Failure

Fig. 4. Generator Re-dispatching

4) Power Supply Reliability Calculation

The power supply reliability is evaluated through LOLP

(Loss of Load Probability) index which can be computed from

(9). LOSS is the total supply interruption period. C is the

number of trial. TIME is the simulation period. The trial is

repeated until 10,000 times so that LOLP is converged to the

steady-state value.

TIMEC

LOSSLOLP

(9)

5) Weekly Fuel Cost Calculation

The weekly fuel cost is calculated at each trial during

Monte Carlo simulation. If the power output of PV and load

demand is constant, the weekly fuel cost is not changed from

the cost which determined by DP. However, the PV output

and the load demand could fluctuate randomly in Monte Carlo

simulation. Therefore, the load dispatch of generators is

changed and its operational cost rises or falls at each trial. In

this paper, the weekly fuel cost is defined as the mean value of

fuel costs of all trials excluding the trial where the generator

failure occurs. This is because the fuel cost increase is the

unexpected factor due to the generator failure. In other words,the outage cost is not determined uniquely.

III. OPTIMAL OPERATION SCHEDULING OF PSHPP

A. Pareto Optimal Solutions

In general, improvement of power system reliability and

reduction of operational cost are competing. Optimization of

the PSHPP scheduling has no unique solution. This type of

problem is a multi-objective optimization problem. In this

paper, pareto optimal solutions are obtained.

8/10/2019 06019279

4/8

4

B. Algorithm Overview

An algorithm overview of the proposed method is shown

in Fig. 5.

Fig. 5. Algorithm Overview

In this study, PSHPP operation schedule is determined by

Genetic Algorithm (GA) and Tabu Search (TS). GA is one of

the evolutionary algorithms, which generates solutions to an

optimization problem using techniques inspired by natural

evolution, such as inheritance, mutation, selection and

crossover. It generates multiple solutions simultaneously at

each generation, and it is good at global search. However GA

is not good at local search. Therefore, this paper proposes that

GA and TS are combined to obtain pareto optimal PSHPP

operation patterns. Tabu search is a local search technique

which enhances the performance using memory structures

named Tabu List. Once a potential solution has been

determined, it is marked as "taboo" so that the algorithm does

not visit that possibility repeatedly. The parameters of thealgorithm are summarized in Table 2.

TABLEII

PARAMETERS OF ALGORITHM

Repeat GenerationsCmax 5000

No. of Genes at Each Genaraion 64

Mutation Rate m 0.05

Repeat Times of Tabu Search 10

No. of Neighborhood Solutions 64

Length of Tabu List 8

GA

TS

Pareto P reservation Strategy

Roulette Selection

Uniform Crossover

Mutation

Selection

Reproduction

C. Algorithm Details

1) Quantization of PSHPP operation pattern

The operation pattern of PSHPP is quantized by 6 steps at

each time t. Figure 6 shows quantization of PSHPP operation

pattern.

OutputofP

SHPP

Time

Pump

Mode

Gene

ratorMode

Stop

100%

-100%

0

75%

50%

25%

5 4 4 3 1 0 0 1 1 5 4

Gene Fig. 6. Quantization of PSHPP Operation Pattern

2) Initial Solutions in GA

The initial genes of the PSHPP operation patterns are

generated by random number at each time period, which are

quantized as mentioned above. If the solution can not satisfy

the capacity constraint of the PSHPP, the solution is rejected

and another solution is generated.

3) Evaluation in GA

Generated genes are evaluated by Power Supply Reliability

Evaluation Model described in section II. The evaluation

indexes are power supply reliability and the fuel cost.

4) Determine the Pareto Optimal Solutions

The pareto optimal solutions are determined from

evaluated solutions. Solutions are labeled by pareto-ranking

method [4]. The ranking is numbered sequentially from

solutions closer to pareto optimal solutions

8/10/2019 06019279

5/8

5

5) Selection and Reproduction in GA

To re-produce the descendants, we use pareto preserve

strategy and uniform crossover. The pareto optimal solutions

are inherited to the next generation as elite whose ranking is

the best. The other descendants are produced by uniform

crossover whose parents are selected by roulette selection.The parent genes are arranged in descending order of their

ranking value. The probabilityPnat which the nth rank gene is

chosen is determined by (10). g is no. of genes at each

generation. Reproduction is implemented through uniform

crossover, whereby each component is randomly selected

from either parent with equal probability. Uniform crossover

has a good characteristic of being unbiased with respect to the

reproduction of genes [5]. Mutation is conducted at the rate m

to prevent the solutions from going into a local optimal

solution.

(10)

6) Tabu Search

TS is conducted whenever GA is repeated 100 times. TS

can optimize only single objective function, TS is repeated

twice from each pareto optimal solutions. First, selecting one

of the pareto solution, TS is carried out from it. The objective

function of TS is LOLP improvement. Second, starting from

the same pareto solution, TS is conducted with objective

function for cost reduction. Figure 7 shows the searching from

pareto optimal solutions. This process is repeated until all

pareto optimal solutions are searched by TS and the pareto

optimal solutions are revised. Then the process goes back tothe GA. It is repeated until the number of generations reaches

Cmax.

Power Supply Reliability

WeeklyFuelCostofThermalPowerPlants

Better

Better

:Tabu Search Direction for LOLP Improvement:Tabu Search Direction for Cost Reduction

Fig. 7. TS Direction from Pareto Optimal Solutions

IV. SIMULATION CONDITION

A. Simulation Model

The simulation is carried out on the modified IEEE 24-bus

Reliability Test System (RTS) [6] with the generator data

shown in the appendix. The system is modified by adding a

300MW PSHPP in addition to the existing hydro power plants.

Since this paper focuses on the power supply reliabilityevaluation, failures in transmission network are not considered.

B. Simulation Condition

Figure 8 depicts two different load conditions for two

seasons, i.e. summer (August) and spring (May). The system

peak load is 3200 MW. In the simulation, the amount of PV

penetration is set to be 0MW and 1000MW.

0

500

1000

1500

2000

2500

3000

Sa t Sun Mon Tue Wed Thu F r i

PowerDemand[MW]

Day

Summer (August)

Spring (May)

Fig. 8. Load Demand Curve

V. SIMULATION RESULTS

A. Pareto Optimal Solutions

The pareto optimal solutions obtained by the proposed

optimization method are shown in Figs. 9 and 10.

PV 0MW

2.9

3

3.1

3.2

3.3

3.4

3.5

3.6

3.7

0 0.05 0.1 0.15 0.2 0.25

WeeklyFuelCo

st[Million$]

LOLP

PV0 MW

PV1000 MW

Fig. 9. Pareto Optimal Solutions (in Summer)

PV 0MW

1.1

1.2

1.3

1.4

1.5

1.6

1.7

0 0.05 0.1 0.15 0.2 0.25

WeeklyFuelCost[Million$]

LOLP

PV0 MW

PV1000 MW

Fig. 10. Pareto Optimal Solutions (in Spring)

g

k

n

k

ngP

1

1

8/10/2019 06019279

6/8

6

In both summer and spring seasons, it can be seen that the

pareto optimal solutions are located in a narrow range of Cost-

LOLP space when PV penetration is 0MW. The pareto

optimal solutions do not include various kinds of operations of

PSHPP because there is no PV output. On the contrary, the

pareto optimal solutions includes various kinds of operationpatterns when PV penetration is 1000MW in both summer and

spring seasons.

B. Operation Schedule of PSHPP

The obtained operation schedule of PSHPP is shown in

Figs. 11 to 14. The best solutions concerning LOLP and the

fuel cost under each condition are discussed here.

0

500

1000

1500

2000

Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperReservoir

[MWh]

Day

Best LOLP Best Cost

Fig. 11. PSHPP Operation Pattern (in Summer with PV 0MW)

0

500

1000

1500

2000

Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperReservoir[MWh]

Day

Best LOLP Best Cost

Fig. 12. PSHPP Operation Pattern (in Summer with PV 1000MW)

0

500

1000

1500

2000

Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperReservoir[MWh

]

Day

Best LOLP Best Cost

Fig. 13. PSHPP Operation Pattern (in Spring with PV 0MW)

0

500

1000

1500

2000

Sa t Sun Mon Tue Wed Thu Fr iCapacityofUpperR

eservoir[MWh]

Day

Best LOLP Best Cost

Fig. 14. PSHPP Operation Pattern (in Spring with PV 1000MW)

In summer, the PSHPP is operated as a pump in the

nighttime and as a generator in the daytime on weekday to

supply the heavy summer load when PV penetration is 0MW.

The role of the PSHPP in this case is load leveling within a

week. It is the conventional role of PSHPP in the powersystem. Compared with the operation patterns between LOLP

best case and cost best case, they are not so much difference.

The near optimal operation pattern is determined uniquely in

summer season when PV penetration is 0MW. In contrast,

when the PV penetration increases to 1000 MW in summer,

the PSHPP is operated as a pump in the daytime and as a

generator in the nighttime in the LOLP best case in spite of no

surplus power. However, PSHPP is hardly operated in the cost

best case.

In spring, the PSHPP is hardly operated when the PV

penetration is 0MW because the operation of the PSHPP does

not contribute much to the fuel cost reduction. When the PV

penetration increases to 1000 MW, PSHPP is hardly operated

in the cost best case, either. The PSHPP has to be operated as

a pump in the daytime both on the weekday and weekend in

the LOLP best case. This pattern is similar to that of summer

pattern.

C. Weekly Generation Schedule

The weekly thermal generation schedules for the PSHPP

operation patterns are shown in Figs. 15 and 16. The cost best

case in summer season when PV penetration is 0MW and the

LOLP best case in spring season when PV penetration is

1000MW are discussed here. In Fig. 15, the PSHPP pushes up

the load demand in the nighttime and pushes it down in the

daytime on weekdays. In Fig. 16, the PSHPP pushes up the

load demand in the daytime and pushes it down in the

nighttime. The deducted load demand shown by solid black

line falls below the minimum output of thermal plants in the

daytime, which shows that the surplus power occurs. However,

in this case, the PSHPP can be operated for both PV output

decrease and increase by emergency operation. It is concluded

that modification of the operation pattern of the PSHPP is

effective for the surplus power problem.

8/10/2019 06019279

7/8

7

0

500

1000

1500

2000

2500

3000

3500

4000

S a t S u n Mo n T u e We d T h u F r i

PowerOutput/LoadDemand[MW]

Day

Maximum Output of Thermal Minimum Output of Thermal

Hydro Nuclear

Load Demand (w/o Pump) Load Demand (with Pump)

Fig. 15. Generation Schedule (in Summer with PV 0MW, for Best Cost)

0

500

1000

1500

2000

2500

3000

3500

4000

S a t S u n Mo n T u e We d T h u F r i

Po

werOutput/LoadDemand[MW]

Day

Maximum Output of Thermal Minimum Output of Thermal

Hydro Nuclear

Load Demand (w/o PV, w/o P ump) Load Demand (with PV, with Pump)

Load Demand (with PV, w/o Pump)

Fig. 16. Generation Schedule (in Spring with PV 1000MW, for Best LOLP)

VI. CONCLUSION

This paper presents a new method for determining an

optimal PSHPP operation pattern which makes it possible to

improve both reliability and economy in the power systems

with a large integration of PV. The simulation results show

that the proposed method is very effective in terms of the

reliability of power system operation. The total fuel cost of

thermal power plants increases in order to secure the power

supply reliability using operation of the PSHPP in power

system with a large penetration of PV. In this paper, the hot

reserve capacity is not considered in scheduling of thermal

power plants. In the future work, it is necessary to develop a

cooperative scheduling method that optimizes both PSHPP

operation and hot reserve capacity of thermal power plants.

VII. APPENDIX

The generator data of IEEE RTS is tabulated in Table 3.

The cost curve and the startup cost of thermal power plants

are shown in Table 4.

TABLEIII

GENERATOR DATA OF IEEERTS(MODIFIED)

1 Thermal (Oil) 5 12 5000 20

2 Thermal (Oil) 4 20 3000 50

3 Hydro 4 50 5000 40

4 Pumped Storage Hydro 1 300 - -

5 Thermal (Coal) 4 76 3926 48

6 Thermal (Oil) 3 100 3926 48

7 Thermal (Coal) 4 155 4110 89

8 Thermal (Oil) 3 197 4110 89

9 Thermal (Coal) 1 350 5541 111

10 Nuclear 2 400 7000 168

Cap. of

each

Unit[MW]

MTBF [h] MTTR [h]Generator

Group Type

No. of

Unit

TABLEIV

THERMAL POWER PLANT DATA

a b c

Thermal (Oil) 5 12 6 16-20 0 .0646 16.8899 49.7108 181.6416

Thermal (Oil) 4 20 10 21-24 0.025 43.5 200 13.356

Thermal (Coal) 4 76 38 6-9 0.0533 9.2374 164.3492 1111.4208

Therma l (Oil) 3 100 50 13-15 0. 0224 19.7128 287.4982 1511.8992

The rma l (Coa l) 4 155 77. 5 2-5 0. 0067 10.2202 207.1786 1777.1544

Therma l (Oil) 3 197 98. 5 10-12 0. 0081 20.2909 378.5918 2070. 18

Thermal (Coal) 1 350 175 1 0.0032 10.102 350.6398 8331.9264

Startup

Cost [$]

Minimum

Output [MW]Type

No. of

Unit

Maximum

Output [MW]

Cost Curve [$/hr]

F(P) = a*P^2+b*P+cMerit

Order

VIII. REFERENCES

[1] R. Aihara, A. Yokoyama, F. Nomiyama and N. Kosugi, Impact of

Operational Scheduling of Pumped Storage Power Plant Considering

Excess Energy and Reduction of Fuel Cost on Power Supply Reliability

in a Power System with a Large Penetration of Photovoltaic

Generations, inProc. 2010 International Conference on Power System

Technology, Hangzhou, China, Oct. 2010.

[2] T. Kuwabara, A. Shibuya, H. Furuta, E. Kita and K. Mitsuhashi, Design

and dynamic response characteristics of 400 MW adjustable speed

pumped storage unit for Ohkawachi Power Station, IEEE Transactions

on Energy Conversion, vol. 11, No. 2, 1996.

[3]

Allen J.Wood and Bruce F.Wollenberg, Power Generation, Operation,and Control, John Wiley& Son, 1983

[4] X. Gandibleux, M. Sevaux, K. Srensen and V. T'Kindt, Metaheuristics

for multiobjective optimization, Springer, 2004.

[5] D. Srinivasan and A.Tettamanzi, Heuristics-guided evolutionary

approach to multiobjective generation scheduling, IEE Proceedings

Generation, Transmission Distribution, vol. 143, No. 6, 1996.

[6] Reliability Test System Task Force of the Application of Probability

Methods subcommittee, IEEE RELIABILITY TEST SYSTEM, IEEE

Transactions on Power Apparatus and Systems, vol. PAS-98, No. 6,

1979.

IX. BIOGRAPHIES

Ryota Aihara was born in Tokyo, Japan, on

September 20, 1984. He received B.S. Eng. fromSophia University, Tokyo, Japan in 2009 and M.S

from The University of Tokyo in 2011. Currently

he is pursuing the Ph.D. degree at The University

of Tokyo. He is a student member of IEEJ.

8/10/2019 06019279

8/8

8

Akihiko Yokoyamawas born in Osaka, Japan, on

October 9, 1956. He received B.S., M.S. and Dr.

Eng. from The University of Tokyo, Tokyo, Japan

in 1979, 1981 and 1984, respectively. He has been

with Department of Electrical Engineering, The

University of Tokyo since 1984 and currently a

professor in charge of Power System Engineering.

He is a member of IEEJ, IEEE and CIGRE.

Fumitoshi Nomiyama was born in 1969. He

received his B.Eng., M.Eng. degrees in electrical

engineering from Doshisha University in 1992 and

1994 respectively. He joined Kyushu Electric

Power Company Inc. in April 1994.

Narifumi Kosugi was born in 1970. He receivedhis B.Eng., M.Eng. degrees in electrical

engineering from Osaka University in 1993 and

1995 respectively. He joined Kyushu Electric

Power Company Inc. in April 1995.