MPEMathematical Problems in Engineering1563-51471024-123XHindawi Publishing Corporation36802410.1155/2009/368024368024Research ArticleA Comparative Study of Three Different Mathematical Methods for Solving the Unit Commitment ProblemKurbanMehmetBaşaran FilikÜmmühanJúdiceJoaquim J.Department of Electrical and Electronics EngineeringAnadolu University, 26470 EskişehirTurkeyanadolu.edu.tr2009206200920092312200810032009100320092009Copyright © 2009This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The unit commitment (UC) problem which is an important subject in power system engineering is solved by using Lagragian relaxation (LR), penalty function (PF), and augmented Lagrangian penalty function (ALPF) methods due to their higher solution quality and faster computational time than metaheuristic approaches. This problem is considered to be a nonlinear programming-(NP-) hard problem because it is nonlinear, mixed-integer, and nonconvex. These three methods used for solving the problem are based on dual optimization techniques. ALPF method which combines the algorithmic aspects of both LR and PF methods is firstly used for solving the UC problem. These methods are compared to each other based on feasible schedule for each stage, feasible cost, dual cost, duality gap, duration time, and number of iterations. The numerical results show that the ALPF method gives the best duality gap, feasible and dual cost instead of worse duration time and the number of iterations. The four-unit Tuncbilek thermal plant which is located in Kutahya region in Turkey is chosen as a test system in this study. The programs used for all the analyses are coded and implemented using general algebraic modeling system (GAMS).

1. Introduction

The UC problem decides to which electricity generation units should be running in each stage to satisfy a predictably varying demand for electricity. The solution of the UC problem is a complex optimization problem, and UC problem can be considered as two linked optimization problems: the first is a combinatorial optimization problem, and the second is a nonlinear programming problem. Due to important startup and shutdown costs, the problem is in general very hard to solve, as it is not possible to perform a separate optimization for each time interval. The exact solution of the UC can be obtained by a complete enumeration of all feasible combination of generating units, which could be very huge number [1, 2].

The problem is structured binary variables making it clearly nonconvex. The binary variables cause a great deal of trouble and reason for the difficulty in solving the UC. Load balance constraint and spinning reserve constraint are coupling constraints across the unit so that one unit affects what will happen on other units if the coupling constraints are met. There have been various methods which are based on mathematical programming and metaheuristic-based for solving the thermal and hydrothermal UC problem in literature. They are based on mathematical programming and metaheuristic based approaches. These major methods are priority list , dynamic programming (DP) , mixed-integer programming , heuristic unit , simulated annealing , tabu search [14, 15], evolutionary programming [16, 17], constraint logic programming , genetic algorithms , LR , interior point method , memetic algorithm , and neural network .

The LR which is one of the most successful approaches for UC is dual optimization technique. This method obtains an appropriate condition to generate feasible solution for UC. One of the most obvious advantages of the LR method is its quantitative measure of the solution quality since the cost of the dual function is a lower bound on the cost of the primal problem . It is seen that LR methods give better solutions than the metaheuristic approaches mentioned above due to both duration time and feasible solution .

The LR, PF, and ALPF methods which present the solving approach based on dual optimization technique are used for solving the UC problem known as an important and hard-solving problem in power system engineering, and the results from the programs coded and implemented using GAMS for these methods are compared to each other. The UC problem for the four-unit Tuncbilek thermal plant is considered for this analysis. A 24-hours day is subdivided into eight discrete stages, and the predicted load of the system can be considered constant over each interval. In this case, the UC problem is an economic combinatorial problem of which solution will yield the minimum operating cost for the scheduling of startup and shutdown of units for each stage in a day.

The remaining sections are outlined as follows. Section 2 provides a description of the UC problem formulation. This section includes an objective function and the constraints of the problem. In Section 3, three methods based on dual optimization techniques are described. In Section 4, the solving algorithm of LR method for the UC problem is explained as a sample. In Section 5, applications and numerical results are presented and discussed. Finally the conclusion is given.

2. UC Problem Formulation

UC has been used to plan over a given time horizon the most economical schedule of committing and dispatching generating units to meet forecasted demand levels and spinning reserve requirements while all generating unit constraints are satisfied. The objective function of the UC problem can be formulated as F(Pit,Uit)=t=1Ti=1N[Fi(Pit)+  SCit(1-Uit-1)]Uit+t=1Ti=1NSDitx(1-Uit-1)Uit-1, where Fi(Pit)  =a(Pit)2+bPit+c is the polynomial function .

The minimization of the objective function is provided to the following constraints.

System Constraints

Power balance constraints. For satisfying the load balance in each stage, the forecasted load demand should be equal to the total power of the generated power for feasible combination: i=1NUitPit=Pdt.Spinning reserve constraints. In order to meet any abnormal operating conditions, the spinning reserve must be considered: Pdt+Prt-i=1NPi,maxUit0.

Unit Constraints

Generation limits. Each unit must satisfy the generation range: Pi,minUitPitPi,maxUit,i=1,2,,N.  Minimum up-time. Once the unit is started up, it should not be shut down before a minimum up-time period is met as follows: (Tion(t-1)-Ti,upt)(Uit-1-Uit)0.Minimum down-time. Once the unit is shut down, there is a minimum down time before it can be started up again: (Tioff(t-1)-Ti,downt)(Uit-Uit-1)0.

3. Methods Based on Dual Optimization Technique

In this paper, three different methods based on dual optimization techniques, LR, PF, and ALPF are used for solving the UC problem. Duality is particularly important in optimization theory. The convergence of the dual optimization methods can be measured by the relative size of the duality gap between the primal and dual solutions. As illustrated in Figure 1, duality gap is obtained by the difference between solutions of primal and dual subproblems. However, primal cost and dual cost functions in reality are not as smooth as the ones in Figure 1 because they are nondifferentiable and nonconvex .

A simplified graphic representation of dual optimization .

General constrained minimization problem is as follows:

PrimalProblem  (P):  Dual  Problem:  (P*),Minimizef(x)maxθ(u)Subjecttog(x)    0u>0,h(x)=0u>0.

The dual procedure attempts to reach the constrained optimum by maximizing the Lagrangian with respect to Lagragian multipliers, while minimizing with respect to the other variables in the problem.

LR Method

LR method can eliminate the dimensionality problem. The LR method is based on the dual optimization theory. The solution of the LR method greatly depends on the Lagragian multipliers. Therefore, setting the initial Lagragian multipliers and updating them are significant to the optimality of the solution. Inappropriate method of updating the Lagrangain multipliers may cause the solution adjustment to oscillate around the global optimum .

The Lagrangian equation is L(x,λ,μ)=f(x)+λigi(x)+μihi(x).

Dual function is q*(λ)=maxq(λ), where q(λ)=min  L(x,λ,μ).

The algorithm of LR method is

Choose starting λ,μ values.

Step 1.

Find a value for each λ and  μ which moves q(λ) toward a larger value.

Step 2.

Assuming that the λ and μ found in Step 1 are now fixed, find the minimum of the L.

PF Method

PF method is one of the widely used methods for obtaining optimal solutions nonconvex problems. It is motivated by the desire to use unconstrained optimization techniques to solve constrained problems. It uses a mathematical function that will increase the objective for any given constrained violation. General transformation of constrained problem into an unconstrained problem  L(x,k,M)=  f0(x)+ki=1n[maks{0,gi(x)}]2+Mj=1m|hj(x)|,f0(x)+ϕ1(g(x),k)+ϕ2(h(x),M).

The algorithm of PF method is

Choose ε  >  0 and β  >1. Choose initial value x1.

Choose penalty factor M1.

Step 3.

Find the following values: Min{f0(x)+  M1[maks{0,g(x)}]},Min{f0(x)  +  Mkαx}=f0(xk)+  Mkαxk.

Step 4.

If Mkαxk<  ε=  0, then stop. Otherwise, Mk+1=βMk,  k  =  k+1. Then go to Step 1.

ALPF Method

The ALPF method is an approach for solving nonlinear programming problems by using the ALPF in a manner that combines the algorithmic aspects of both LR and PF methods .

Lagragian equation is L(x,c,v,M)=  f0(x)+i=1mvigi(x)+i=1qvjhj(x)+Mi[i=1mgi(x)]2+Mj[i=1qhj(x)2].

The algorithm of ALPF method is as follows.

Initialization

Select some initial Lagrangian multipliers. Lagrangian multipliers are vm and Mm. These are can be written as follows: vm=(v11,v21,,vm1),Mm=(M11,M21,,Mm1), where vm is positive values, and Mm is for the penalty parameters.

Let x0 be a null vector and denote VIOL(x0)=+, for any x En.

Measure the level of the violation by using (3.7)

VIOL(x)=max1im{|fi(x)|}.

Choose k=1;  ε>0 is a measure of constraint violations. Put k  =  1, and proceed to the “inner loop’’ of the algorithm.

Inner Loop

Solve minimum of the Lagrangian subject toxEn, and let xk be the optimal solution obtained. If VIOL(xk)  =0 or VIOL(xk)<ε, then stop. Otherwise, if VIOL(xk)<(1/4)VIOL(xk-1), then go to outer loop. If VIOL(xk)>VIOL(xk-1), then, for each constraint i=  1,,l for which fi(xk)>(1/4)VIOL(xk-1), replace the corresponding penalty parameter Mi by 10 Mi and repeat this inner loop step.

Outer Loop

Replace v by vnew, where vik+1=vik+2Mikfi(xk). Increment k by 1, and return to the inner loop.

4. Solving Algorithm for UC

The LR method solves the UC problem by relaxing or temporarily ignoring the coupling constraints. This is done through the dual optimization procedure attempting to reach the constrained optimum by maximizing the Lagrangian. In this section, the solving algorithm of LR method for the UC problem is explained as a sample. The Lagrangian equation can be written as follows : L(P,U,λ)=t=1Ti=1NF(Pit,Uit)+SCit(1-Uit-1)Uit+t=1Ti=1NSDitx(1-Uit-1)+t=1Tλt(Ploadt-i=1NPitUit)+t=1Tμt(Ploadt+Pr(t)-i=1NPi,maxUit).

q*(λ,μ)   is a dual function, and it can be shown as follows:

q*(λ,μ)=  maxλt,μt  q(λ,μ), where q(λ,μ)=  min  Pit,Ui,tL(P,U,λ,μ),L=  t=1Ti=1N[F(Pit,Uit)  +  SCit(1-Uit-1)Uit]+t=1Ti=1NSDitx(1-Uit-1)+t=1TλtPloadt+μt(Ploadt+Pr(t))-i=1NλtPitUit-μtPi,maxUit, where λtPloadt, and μt  (Ploadt+Pr(t)) are constant and can be dropped.

Goal of separating the units from one another has been achieved. The term inside the outer brackets t=1T{[F(Pit,Uit)+SCit(1-Uit-1)Uit+t=1Ti=1NSDitx(1-Uit-1)]}-  λtPitUit-1-μtPi,maxUi,t.

Equation (4.4) can be solved separately for each generating unit, without regard for what is happening on the other generating units. The minimum of the Lagrangian is found by solving for the minimum for each generating unit over all time periods

minq(λ)=i=1Nmin  t=1T{[F(Pit,Uit)+  SCit(1-Uit-1)Uit+t=1Ti=1NSDitx(1-Uit-1)]}-λtPitUit-μtPi,maxUit-1.

This is solved as a one-variable DP problem. This can be visualized in Figure 2, which shows the only two possible states for unit i.

Only two possible states for unit i (Uit=0 or Uit=1), where Si is the startup cost for unit i.

At the Ui,t=0 state, the value of the function to be minimized is trivial. At the state Ui,t=1, the function to be minimized is.

min[FiPit-λtPit-μtPi,max].

The minimum of this function is

ddPit[FiPit-λtPit-μtPi,max]  =0.

The solution to this equation is

dFi(Pit,opt)dPit  =  λt,  Pit,opt  =  λt-bi2ci.

There are three cases in this situation.

IfPit,opt<  P,  i,min    then  Pit=Pi,min.

IfP  i,min  <  Pit,opt<  Pi,max,  thenPit=  Pit,opt.

IfPit,opt>  P,  i,max    thenPit=  P.  i,max

DP is used to determine the optimal schedule of each unit over the scheduled time period. To minimize F(Pit)-λtPit<0 at each stage and that when Uit=0, this value goes to zero, then the only way to get a value lower is to have 

F(Pit)-λtPit<0.

In the LR method, the value obtained by (2.1) and problem constraints is called the primal value (J*) of the problem. The value in (3.3) represents the dual value. Whether the problem converges to its solution is decided by looking at the relative duality gap value. This value is calculated using (4.10) Relative  duality  gap=  J*-q*q*.

5. Applications and Numerical Results

The LR, PF, and ALPF methods are applied for solving the UC problem for four-unit Tuncbilek thermal plant separately, and the numerical results are given in tables. These methods are compared to each other based on feasible schedule for each stage, feasible cost, dual cost, duality gap, duration time, and number of iterations using the table in the comparative results part. The spinning reserve demand is set as 10% of the system demand. All these methods are coded in GAMS which is a high-level modeling system for mathematical programming problems  and run on Pentium 4—3 GHz personal computer. CONOPT is used as a GAMS solver. The data used in the analysis is taken from Turkish Electric Power Company and Electricity Generation Company. The unit characteristics for four-unit Tuncbilek thermal plant are given in Table 1.

Unit characteristics for four-unit Tuncbilek thermal plant.

UnitPmin  (MW)Pmax (MW)SC (\$)a (\$/h)b (\$/MWh)c (\$/MW2h)SD (\$)Tup (h)Tdown  (h)
1832600.51510.86149.912011
217652400.2278.341284.648022
3351505500.0829.9441495.8110033
4301505500.07412.44388.9110033

In this study, a 24-hours day is subdivided into 8 discrete stages. The load demands for the stages are given in Table 2.

11685313
21506347
32607308
42758231
For LR Method

LR method is applied to UC problem. The feasible unit combination for the LR method is given in Table 3. The initial of the λ and μ values is chosen as zero.

Feasible unit combination for LR method.

StageUnit combination
10 0 1 1
20 0 1 1
30 1 1 1
41 1 1 1
51 1 1 1
61 1 1 1
71 1 1 1
80 1 1 1
For PF Method

PF method is applied to UC problem. The feasible unit combination for the PF method is given in Table 4. The initial of the k and M values are 320 and 970, respectively.

Feasible unit combination for PF method.

StageUnit combination
10 0 1 1
20 0 1 1
30 1 1 1
40 1 1 1
51 1 1 1
61 1 1 1
70 1 1 1
80 1 1 1
For ALPF Method

ALPF method is applied to UC problem. The feasible unit combination for the ALPF method is given in Table 5. The initial of the v and M values is 25 and 100, respectively.

Feasible unit combination for ALPF method.

StageUnit combination
10 0 1 1
20 0 1 1
30 1 1 1
40 1 1 1
50 1 1 1
61 1 1 1
70 1 1 1
80 1 1 1
Comparative Results

Feasible cost, dual cost and duality gap values, duration time, and number of iterations are found for each method. These numerical results are given in Table 6.

The numerical results for LR, PF, and ALPF methods.

Compared itemsLRPFALPF
Feasible cost (\$)56381.93155801.47755233.897
Dual cost (\$)55428.01455305.33254894.266
Duality gap (%)1.17210.89710.6187
Number of iterations193667
Duration time (s)5.323.5614.63

It can be seen from Table 6 that LR has the highest duality gap value, and ALPF performs much better than LR and PF for duality gap value. It is noticed that ALPF method yields the minimum duality gap value with minimum cost. On the other hand, LR method reaches the solution with the minimum number of iterations. In addition to this, PF method yields the fastest solution.

6. Conclusion

Three different dual approach-based methods, LR, PF and ALPF methods, are used for solving the UC problem known as an important and hard-solving problem in power system engineering, and the results from the programs coded and implemented using GAMS for these methods are compared to each other according to feasible cost, dual cost, duality gap, duration time, and number of iterations. The numerical results show that the ALPF method which is firstly used for solving the UC problem in literature gives the best duality gap, feasible and dual cost values instead of worse duration time and the number of iterations. It is seen that ALPF method yields the minimum duality gap value with minimum cost. On the other hand, LR method reaches the solution with the minimum number of iterations. In addition to this, PF method yields the fastest solution. Since the most important factor in the solution of UC problems is feasible cost value, it can be said that ALPF is the most suitable method among others.

List of symbolsf(x):

Objective function

H̅:

Upper bound of the dual function

h(x):

Equality constraint

g(x):

Inequality constraint

k:

Multiplier for PF method

M:

Multiplier for PF and ALPF method

N:

Number of generating units

P:

Primal problem

P*:

Dual problem

Pdt:

Nominal demand at hour t (MW)

Pit:

Generation output of unit i at hour t (MW)

Pmax,it:

Maximum available capacity of unit i at hour t (MW)

Pmin,it:

Minimum available capacity of unit i at hour t (MW)

Prt:

System spinning reserve at hour t (MW)

RUi:

Ramp up rate of unit i (MW/h)

RDi:

Ramp down rate of unit i (MW/h)

SCi:

Start up cost of unit i  (\$)

SDi:

Shut down cost of unit i  (\$)

Uit:

Status value of unit i at time t

T:

Time horizon for UC (h)

Ti,down:

Minimum down-time of unit i,  h

Ti,offt:

Continuously off-time of unit i up to hour t,  h

Ti,ont:

Continuously on-time of unit i up to hour t,  h

Ti,down:

Minimum up-time of unit i,  h

TC:

Total cost (\$)

v:

Multiplier for ALPF.

MantawyA. H.Abdel-MagidY. L.SelimS. Z.A simulated annealing algorithm for unit commitmentIEEE Transactions on Power Systems199813119720410.1109/59.651636VianaA.SousaJ. P.MatosM.Simulated annealing for the unit commitment problem2Proceedings of IEEE Porto Power Tech ConferenceSeptember 2001Porto, Portugal1410.1109/PTC.2001.964747LeeF. N.FengQ.Multi-area unit commitmentIEEE Transactions on Power Systems19927259159910.1109/59.141764LeeF. N.The application of commitment utilization factor (CUF) to thermal unit commitmentIEEE Transactions on Power Systems19916269169810.1109/59.76714BurnsR. M.GibsonC. A.Optimization of priority lists for a unit commitment programProceedings of IEEE/PES Summer MeetingJuly 1975San Francisco, Calif, USApaper A 75 453-1ChenC.-L.ChenN.Strategies to improve the dynamic programming for unit commitment applicationTransactions of the Chinese Institute of Engineers200293181189ArroyoJ. M.ConejoA. J.Optimal response of a thermal unit to an electricity spot marketIEEE Transactions on Power Systems20001531098110410.1109/59.871739GuanX.ZhaiQ.PapalexopoulosA.Optimization based methods for unit commitment: Lagrangian relaxation versus general mixed integer programming2Proceedings of IEEE/PES Summer MeetingJuly 2003Toronto, Canada1095110010.1109/PES.2003.1270468LiT.ShahidehpourM.Price-based unit commitment: a case of Lagrangian relaxation versus mixed integer programmingIEEE Transactions on Power Systems20052042015202510.1109/TPWRS.2005.857391AnnakkageU. D.NummondaT.PahalawaththaN. C.Unit commitment by parallel simulated annealingIEE Proceedings: Generation, Transmission and Distribution1995142659560010.1049/ip-gtd:19952215NayakR.SharmaJ. D.Hybrid neural network and simulated annealing approach to the unit commitment problemComputers and Electrical Engineering200026646147710.1016/S0045-7906(99)00037-3RajanC. C. A.MohanM. R.ManivannanK.Refined simulated annealing method for solving unit commitment problem1Proceedings of the International Joint Conference on Neural Networks (IJCNN '02)May 2002Honolulu, Hawaii, USA33333810.1109/IJCNN.2002.1005493MantawyA. H.Abdel-MagidY. L.SelimS. Z.A simulated annealing algorithm for unit commitmentIEEE Transactions on Power Systems199813119720410.1109/59.651636RajanC. C. A.MohanM. R.ManivannanK.Neural based Tabu search method for solving unit commitment problemProceedings of the 5th International Conference on Power System Management and ControlApril 2002London, UK180185MoriH.UsamiT.Unit commitment using Tabu search with restricted neighborhoodProceedings of the International Conference on Intelligent Systems Applications to Power Systems (ISAP '96)January-February 1996Orlando, Fla, USA422427RajanC. C. A.MohanM. R.An evolutionary programming-based Tabu search method for solving the unit commitment problemIEEE Transactions on Power Systems200419157758510.1109/TPWRS.2003.821472JusteK. A.KitaH.TanakaE.HasegawaJ.An evolutionary programming solution to the unit commitment problemIEEE Transactions on Power Systems19991441452145910.1109/59.801925HuangK.-Y.YangH.-T.HuangC.-L.A new thermal unit commitment approach using constraint logic programmingIEEE Transactions on Power Systems1998133936945SwarupK. S.YamashiroS.Unit commitment solution methodology using genetic algorithmIEEE Transactions on Power Systems2002171879110.1109/59.982197PadhyN. P.Unit commitment using hybrid models: a comparative study for dynamic programming, expert system, fuzzy system and genetic algorithmsInternational Journal of Electrical Power & Energy Systems200123882783610.1016/S0142-0615(00)00090-9SenjyuT.YamashiroH.UezatoK.FunabashiT.A unit commitment problem by using genetic algorithm based on unit characteristic classification1Proceedings of IEEE/PES Winter MeetingJanuary 2002Singapore586310.1109/PESW.2002.984954YangP.-C.YangH.-T.HuangC.-L.Solving the unit commitment problem with a genetic algorithm through a constraint satisfaction techniqueElectric Power Systems Research1996371556510.1016/0378-7796(96)01036-XKazarlisS. A.BakirtzisA. G.PetridisV.A genetic algorithm solution to the unit commitment problemIEEE Transactions on Power Systems19961118392NievaR.IndaA.GuillenI.Lagrangian reduction of search-range for large-scale unit commitmentIEEE Transactions on Power Systems198722465473VirmaniS.AdrianE. C.ImhofK.MukherjeeS.Implementation of a Lagrangian relaxation based unit commitment problemIEEE Transactions on Power Systems1989441373138010.1109/59.41687BatutJ.RenaudA.Daily generation scheduling optimization with transmission constraints: a new class of algorithmsIEEE Transactions on Power Systems19927398298910.1109/59.207311WangS. J.ShahidehpourS. M.KirschenD. S.MokhtariS.IrisarriG. D.Short-term generation scheduling with transmission and environmental constraints using an augmented Lagrangian relaxationIEEE Transactions on Power Systems19951031294130110.1109/59.466524RedondoN. J.ConejoA. J.Short-term hydro-thermal coordination by Lagrangian relaxation: solution of the dual problemIEEE Transactions on Power Systems1999141899510.1109/59.744490ZhaiQ.GuanX.CuiJ.Unit commitment with identical units: successive subproblem solving method based on Lagrangian relaxationIEEE Transactions on Power Systems20021741250125710.1109/TPWRS.2002.805003OngsakulW.ongsakul@ait.ac.thPetcharaksN.Unit commitment by enhanced adaptive Lagrangian relaxationIEEE Transactions on Power Systems200419162062810.1109/TPWRS.2003.820707SriyanyongP.SongY. H.Unit commitment using particle swarm optimization combined with Lagrange relaxation3Proceedings of IEEE/PES General MeetingJune 2005San Francisco, Calif, USA2752275910.1109/PES.2005.1489390FuY.ShahidehpourM.LiZ.Long-term security-constrained unit commitment: hybrid Dantzig-Wolfe decomposition and subgradient approachIEEE Transactions on Power Systems20052042093210610.1109/TPWRS.2005.857286KimballL. M.ClementsK. A.DavisP. W.NejdawiI.Multiperiod hydrothermal economic dispatch by an interior point methodMathematical Problems in Engineering200281334210.1080/10241230211378MR1918088ValenzuelaJ.SmithA. E.A seeded memetic algorithm for large unit commitment problemsJournal of Heuristics20028217319510.1023/A:1017960507177SasakiH.WatanabeM.KubokawaJ.YorinoN.YokoyamaR.A solution method of unit commitment by artificial neural networksIEEE Transactions on Power Systems19927397498110.1109/59.207310LiangR.-H.KangF.-C.Thermal generating unit commitment using an extended mean field annealing neural networkIEE Proceedings: Generation, Transmission and Distribution2000147316417010.1049/ip-gtd:20000303HuangS.-J.HuangC.-L.Application of genetic-based neural networks to thermal unit commitmentIEEE Transactions on Power Systems1997122654660DieuV. N.OngsakulW.Improved merit order and augmented Lagrange Hopfield network for unit commitmentIET Generation, Transmission & Distribution20071454855610.1049/iet-gtd:20060321BazaraaM. S.SheraliH. D.ShettyC. M.Nonlinear Programming: Theory and Algorithms19932ndHoboken, NJ, USAJohn Wiley & SonsMR2218478ZBL0774.90075WoodA. J.WollenbergB. F.Power Generation Operation and Control19962ndNew York, NY, USAWiley-InterscienceBrookeA.KendrickD.MeerausA.RamanR.GAMS: a user's guideGAMS Development Corporation, 1998, http://www.gams.com/