Julia-带有@expression的表达式集合必须是线性的

我正在研究最小二乘法问题，并尝试使用跳跃和古罗比将给定的n个点拟合成2条曲线

此代码用于创建一条曲线

using JuMP, Gurobi
m = Model(solver=GurobiSolver(OutputFlag=0))
@variable( m, x[1:size(A,2)] )
@objective( m, Min, sum((A*x-b).^2) )
solve(m)

摘自本幻灯片

我的尝试就在这里

# define (x,y) coordinates of the points
x = [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 2 ]
y = [ 1, 3, 0, 1, 2, 4, 6, 7, 5, 4 ]

x1 = x[1:5]
x2 = x[5:9]

y1 = y[1:5]
y2 = y[5:9]
using PyPlot
figure(figsize=(8,4))
plot(x,y,"r.", markersize=10)
axis([0,10,-2,8])
grid("on")

println(x1)
println(x2)
println(y1)
println(y2)

# order of polynomial to use
k = 3

# fit using a function of the form f(x) = u1 x^k + u2 x^(k-1) + ... + uk x + u{k+1}
n = length(x1)

A = zeros(n,k+1)
for i = 1:n
    for j = 1:k+1
        A[i,j] = x1[i]^(k+1-j)
    end
end

A1 = zeros(n,k+1)
for i = 1:n
    for j = 1:k+1
        A1[i,j] = x2[i]^(k+1-j)
    end
end

println(A)
print(A1)


# NOTE: must have either Gurobi or Mosek installed!

using JuMP, Gurobi

#m = Model(solver=MosekSolver(LOG=0))
m = Model(with_optimizer(Gurobi.Optimizer))
#m = Model(solver=GurobiSolver(OutputFlag=1,NumericFocus=2))    # extra option to do extra numerical conditioning
#m = Model(solver=GurobiSolver(OutputFlag=1,BarHomogeneous=1))  # extra option to use alternative algorithms

@variable(m, u[1:k+1])
@variable(m, v[1:k+1])
@variable(m, 0 <= s[1:k+1] <= 1,integer=true)

@expression(m, su[i in 1:k+1] , u'.*A[i,:])
@expression(m, sv[i in 1:k+1] , v'.*A[i,:])

#@expression(model, q_expr[k=1:2], sum([sum([x[i]*x[j]*c[k] for i in 1:3]) for j in 1:3]))


@objective(m, Min, sum( (1-s[i])*(y[i]-su[i]).^2 + s[i]*(y[i]-sv[i]).^2 for i in 1:k+1 ) )

optimize!(m)
uopt = value.(u)
println(uopt)

---

这没有帮助

这似乎基本上是同一个问题。如果我正确地理解了你想要的，那么你试图解决的问题不是二次问题，所以你将无法使用Gurobi来处理它。这看起来基本上和你的问题相同。如果我正确地理解了您想要的，那么您试图解决的问题不是二次问题，因此您将无法使用Gurobi处理它。

Collection of expressions with @expression must be linear. For quadratic expressions, use your own array.