F = L2 + I2 + 0.5*B1*u1²+ 0.5*B2*u2²

函数[zopt,cost,exitflag] = ResistantTBSolver(FUN,z0,A,b,Aeq,beq,lb,ub,NONLINCON)这是一个调用XYZ非线性编程的接口函数%求解器来解决nlmpc控制器定义的NLP问题。的该函数的输入和输出定义与fmincon相同。％这个接口函数将fmincon输入转换为所需的格式%，并将结果转换回fmincon输出。％%的输入% FUN:非线性成本。FUN接受输入z(决策变量)和%返回成本f(在z处计算的标量值)及其%梯度g(一个nz-by-1向量在z处的值)。% z0: z的初始猜测% A,b: A*z<=b% Aeq,beq: Aeq*z==beq% lb,ub: z的下界和上界NONLINCON:非线性约束。NONLINCON接受输入z并返回%，向量C和Ceq作为前两个输出，表示其中，非线性不等式和不等式分别为% FUN在C(z) <= 0和Ceq(z) = 0时最小化。NONLINCON也返回C (a nz-by-ncin)的雅可比矩阵%稀疏矩阵)和Ceq (nz-by-nceq稀疏矩阵)。％%输出% zopt: z的最优解%成本:在最优z处的最优成本% exitflag:满足一阶最优性条件。% 0太多的函数计算或迭代。% -1由输出/绘图函数停止。% -2未找到可行点。％您可以使用此示例作为模板来实现接口文件%你自己的NLP求解器。的M文件或MEX文件% MATLAB路径。％%不编辑以上行。％The MathWorks, Inc.版权所有设置线性和非线性约束的维度信息num_lin_ineq = size(A,1);num_lin_eq = size(Aeq,1);[in,eq] = NONLINCON(z0);Num_non_ineq = length(in);Num_non_eq = length(eq);Total = num_non_ineq + num_non_eq + num_lin_ineq + num_lin_eq;Logicals_nlineq = false(total,1);Logicals_nlineq (1:num_non_ineq) = true;Logicals_nleq = false(total,1);Logicals_nleq (num_non_ineq+(1:num_non_eq)) = true; logicals_ineq = false(total,1); logicals_ineq(num_non_ineq+num_non_eq+(1:num_lin_ineq)) = true; logicals_eq = false(total,1); logicals_eq(num_non_ineq+num_non_eq+num_lin_ineq+(1:num_lin_eq)) = true; options = struct(“nlineq”logicals_nlineq,“nleq”logicals_nleq,.．.“ineq”logicals_ineq,“情商”, logicals_eq);设置决策变量边界选项。Lb = Lb;选项。Ub = Ub;设置非线性约束的RHS选项。Cl = [-inf(num_non_ineq,1); 0 (num_non_eq,1)];选项。Cu = [0 (num_non_ineq,1); 0 (num_non_eq,1)];设置线性约束的RHS选项。Rl = [-inf(num_lin_ineq,1);beq];Options.ru = [b;beq];设置一个矩阵选项。一个＝sparse([A; Aeq]);设置XYZ求解器选项选项。算法= struct(“print_level”0,“max_iter”, 200,.．.“max_cpu_time”, 1000,“托尔”1.0000 e-06.．.“hessian_approximation”，内存有限的）;设置XYZ求解器使用的函数句柄Jstr = sparse(ones(num_non_ineq+num_non_eq,length(z0)));Funcs = struct(“目标”@ (x) fval(有趣,x).．.“梯度”@ (x) gval(有趣,x).．.“约束”@ (x) conval (NONLINCON x),.．.的雅可比矩阵@ (x) jacval (NONLINCON x),.．.“jacobianstructure”Jstr @ ().．.）;调用XYZ并返回成本和状态[zopt,output] = XYZsolver(z0,funcs,options);cost = FUN(zopt);exitflag = convertStatustoExitflag(output.status);实用函数函数F = fval(fun,z)返回非线性成本[f，~] = fun(z);函数G = gval(fun,z)%回报成本梯度[~，g] = fun(z);函数C = conval(非lcon,z)返回非线性约束[in,eq] = nonlcon(z);C = [in;eq];函数J = jacval(非lcon,z)在稀疏矩阵中返回约束雅可比矩阵为nc × nzJin是nz-by-ncin稀疏的，Jeq是nz-by-nceq稀疏的[~，~，Jin,Jeq] = nonlcon(z);J =[金杰克]';函数exitflag = convertStatustoExitflag(状态)开关(状态)情况下0%的信息。年代tatus = 'Success';Exitflag = 1;情况下1%的信息。年代tatus = 'Solved to Acceptable Level';Exitflag = 1;情况下2%的信息。年代tatus = 'Infeasible';Exitflag = -1;情况下3.%的信息。年代tatus = 'Search Direction Becomes Too Small';Exitflag = -2;情况下4%的信息。年代tatus = 'Diverging Iterates';Exitflag = -2;情况下6%的信息。年代tatus = 'Feasible Point Found';Exitflag = 1;情况下－1%的信息。年代tatus = 'Exceeded Iterations';Exitflag = 0;情况下4%的信息。年代tatus = 'Max Time Exceeded';Exitflag = 0;否则%的信息。年代tatus = 'IPOPT Error';Exitflag = -2;结束