# tunning an RHMS model a function for tunning an RHMS model based on a set of observed time series, using **particle swarm optimization** ```r tune(object,targetObject,decisionObjects, observationTS,delay=0, transformBandWith=list(ct=c(1 , 2.5), cp=c(0.1, 0.3), cn=c(25 , 85 ), k =c(0.1, 2 )), routingBandWith=list(manning = c(0.0001, 0.1), x = c(0.2 , 0.6), k = c(1 , 5 )), maxiter=NA,update=FALSE,plot=FALSE) ``` ## Arguments - `object`: an object from class of `createBasin` - `targetObject`: an object from either of classes: `createDiversion`, `createReservoir`, `createSubbasin`, `createJunction`, `createReach` associated to the `observationTS` - `decisionObjects`: A list of objects, also, already existing in the `object` which their parameters needed to be optimized. They objects must be from either of classes: `createSubbasin`, `createReach` - `observationTS`: a vector: an observed flow time series (cms) - `delay`: (optional) an integer presenting the number of time steps to delay `observationTS` time series - `transformBandWith`: an list: a list of vector(s), including upper and lower limit of parameters of tansformation methods. Each parameter search domain is set as a two-value vector, whose first element indicates lower limit and second elemnt is upper limit. * `Ct`=[1, 2.5] and `Cp`=[0.1, 0.3] are parameters for `"Snyder"` Unit Hydrograph (SUH) * `cn`=[25, 85] curve number for `"SCS"` loss method * `k` for `"horton"` loss method - `routingBandWith`: an list: a list of vector(s), including upper and lower limit of parameters of routing methods. Each parameter search domain is set as a two-value vector, whose first element indicates lower limit and second elemnt is upper limit. * `manning`=[0.0001, 0.1] is a parameter used `"muskingumcunge"` method * `x` = [0.2, 0.6] and `k`=[1, 5] belong to `"muskingum"` channel routing method - `maxiter`: (optional) an integer: maximum number of iterations. default to the square of dimension of decision variables - `plot`: (optional) logical: plots the optimization results - `update`: (optional) logical: If FALSE, the optimized parameter(s) are returned,If TRUE, the calibrated object from class of `createBasin` is returned ## Returns a vector of tunned parameters or an object from class of `createBasin` ## References Kennedy, J. (1997). "The particle swarm: social adaptation of knowledge". Proceedings of IEEE International Conference on Evolutionary Computation. pp. 303-308 ## Examples ```r J1<-createJunction (name="J1") R1<-createReach(name="R1",routingMethod="muskingum", routingParams=list(k=3,x=0.2), downstream=J1) R2<-createReach(name="R2",routingMethod="muskingumcunge", routingParams=list(bedWith=50, sideSlope=2, channelSlope=0.0005, manningRoughness=0.025, riverLength=100), downstream=J1) S1<-createSubbasin(name = "S1", precipitation=sin(seq(0,pi,length.out=20))*40, Area=100,downstream=R1, transformMethod="SCS",lossMethod="SCS", transformParams=list(Tlag=4),lossParams=list(CN=60)) S2<-createSubbasin(name = "S2", precipitation=sin(seq(0,pi,length.out=20))*30, Area=300,downstream=R2, transformMethod="snyder",lossMethod="horton", transformParams=list(Cp=0.17,Ct=2,L=30,Lc=15), lossParams=list(f0=10,f1=4,k=1)) basin1<-createBasin(name = "Ghezil_Ozan", simulation=list(start='2000-01-01', end ='2000-01-05', by =3600)) basin1<-addObjectToBasin(S1, basin1) basin1<-addObjectToBasin(S2, basin1) basin1<-addObjectToBasin(R1, basin1) basin1<-addObjectToBasin(R2, basin1) basin1<-addObjectToBasin(J1, basin1) ## Not run: plot(basin1) simulated<-sim(basin1) plot(simulated) observationTS1<-simulated$operation$junctions[[1]]$outflo[,1] set.seed(1) observationTS1<-observationTS1+rnorm(length(observationTS1),0,25) y<-observationTS1; x<-1:length(observationTS1) observationTS1<-predict(loess(y~x),x) observationTS1[which(observationTS1<0)]<-0 observationTS<-observationTS1 plot(simulated$operation$junctions[[1]]$outflow[,1],typ='o',ylab='flow rate (cms)',xlab='time step') lines(observationTS,col=2) transformBandWith=list(ct=c(1 ,2.5), cp=c(0.1,0.3), cn=c(25 ,85) , k =c(0.1,2)) routingBandWith=list(maning = c(0.0001,0.1), x = c(0.2 ,0.6), k = c(1 ,5)) targetObject<-J1 decisionObjects<-list(R1,R2,S1,S2) ## Not run: tune(object=basin1, targetObject=targetObject, decisionObjects=decisionObjects, observationTS=observationTS, routingBandWith=routingBandWith, transformBandWith=transformBandWith, plot=TRUE) ## End(Not run) ``` ## Author(s) Rezgar Arabzadeh

tune function