% ------------------------------------------------------------------------------ % Subroutine for the derivative of Jac with respect to rate coefficients % Times a user vector % ----------------------------------------------------------------------------- DJDR = function dJac_dRcoeff( V, F, U, NCOEFF, JCOEFF ) % V - Concentrations of variable/fixed species % KPP_REAL V(KPP_NVAR), F(NFIX) % U - User-supplied Vector % KPP_REAL U(KPP_NVAR) % NCOEFF - the number of rate coefficients with respect to which we differentiate % INTEGER NCOEFF % JCOEFF - a vector of integers containing the indices of reactions (rate % coefficients) with respect to which we differentiate % INTEGER JCOEFF(NCOEFF) % DFDR - a matrix containg derivative values; specifiy, % column j contains d Jac(1:KPP_NVAR) / d RCT( JCOEFF(j) ) * U % for each 1 <= j <= NCOEFF % This matrix is stored in a column-wise linearized format % KPP_REAL DJDR(KPP_NVAR*NCOEFF) % Local vector for Jacobian of reactant products % KPP_REAL JV_RPROD(NJVRP) % Compute the Jacobian of all reactant products JV_RPROD = JacReactantProd( V, F ); % Compute the derivatives by multiplying column JCOEFF(j) of the stoichiometric matrix with A_PROD for j=1:NCOEFF % Initialize the j-th column of derivative matrix to zero for i=1,KPP_NVAR DJDR(i+KPP_NVAR*(j-1)) = 0.0; end % Column JCOEFF(j) in the stoichiometric matrix times the % ( Gradient of reactant product of the JCOEFF(j)-th reaction X user vector ) % give the j-th column of the derivative matrix % % Row JCOEFF(j) of JV_RPROD times the user vector aj = 0.0; for k=CROW_JVRP(JCOEFF(j)):CROW_JVRP(JCOEFF(j)+1)-1 aj = aj + JV_RPROD(k)*U(ICOL_JVRP(k)); end % Column JCOEFF(j) of Stoichiom. matrix times aj for k=CCOL_STOICM(JCOEFF(j)):CCOL_STOICM(JCOEFF(j)+1)-1 DJDR(IROW_STOICM(k)+KPP_NVAR*(j-1)) = STOICM(k)*aj; end end return % dJac_dRcoeff