how can i solve this problem?

21 次查看（过去 30 天）
Emad Samir
Emad Samir 2021-12-23
回答： John D'Errico 2021-12-23
syms x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 y1 y2 y3 y4 y5 y6 y7 y8 y9 y10;
eq1=y1*(30*x2 + 40*x3 + 50*x4 + 60*x5 + 70*x6 + 30*x7 + 40*x8 + 50*x9 + 60*x10)-550;
eq2=y2*(50*x1 + 40*x3 + 50*x4 + 30*x5 + 50*x6 + 40*x7 + 50*x8 + 60*x9 + 30*x10)-600;
eq3=y3*(50*x1 + 30*x2 + 60*x4 + 20*x5 + 50*x6 + 40*x7 + 50*x8 + 40*x9 + 20*x10)-500;
eq4=y4*(50*x1 + 40*x2 + 80*x3 + 50*x5 + 60*x6 + 50*x7 + 30*x8 + 40*x9 + 50*x10)-600;
eq5=y5*(30*x1 + 50*x2 + 50*x3 + 60*x4 + 60*x6 + 50*x7 + 30*x8 + 30*x9 + 60*x10)-700;
eq6=y6*(40*x1 + 50*x2 + 50*x3 + 50*x4 + 40*x5 + 60*x7 + 40*x8 + 50*x9 + 70*x10)-700;
eq7=y7*(40*x1 + 50*x2 + 40*x3 + 50*x4 + 30*x5 + 80*x6 + 40*x8 + 50*x9 + 30*x10)-600;
eq8=y8*(30*x1 + 30*x2 + 40*x3 + 40*x4 + 20*x5 + 30*x6 + 90*x7 + 40*x9 + 30*x10)-500;
eq9=y9*(50*x1 + 40*x2 + 80*x3 + 40*x4 + 20*x5 + 30*x6 + 30*x7 + 40*x8 + 40*x10)-550;
eq10=y10*(50*x1 + 40*x2 + 40*x3 + 30*x4 + 50*x5 + 40*x6 + 20*x7 + 50*x8 + 30*x9)-600;
eq11=x1*(30*y2 + 40*y3 + 50*y4 + 60*y5 + 70*y6 + 30*y7 + 40*y8 + 50*y9 + 60*y10)-600;
eq12=x2*(50*y1 + 40*y3 + 50*y4 + 30*y5 + 50*y6 + 40*y7 + 50*y8 + 60*y9 + 30*y10)-700;
eq13=x3*(50*y1 + 30*y2 + 60*y4 + 20*y5 + 50*y6 + 40*y7 + 50*y8 + 40*y9 + 20*y10)-600;
eq14=x4*(50*y1 + 40*y2 + 80*y3 + 50*y5 + 60*y6 + 50*y7 + 30*y8 + 40*y9 + 50*y10)-700;
eq15=x5*(30*y1 + 50*y2 + 50*y3 + 60*y4 + 60*y6 + 50*y7 + 30*y8 + 30*y9 + 60*y10)-500;
eq16=x6*(40*y1 + 50*y2 + 50*y3 + 50*y4 + 40*y5 + 60*y7 + 40*y8 + 50*y9 + 70*y10)-700;
eq17=x7*(40*y1 + 50*y2 + 40*y3 + 50*y4 + 30*y5 + 80*y6 + 40*y8 + 50*y9 + 30*y10)-550;
eq18=x8*(30*y1 + 30*y2 + 40*y3 + 40*y4 + 20*y5 + 30*y6 + 90*y7 + 40*y9 + 30*y10)-500;
eq19=x9*(50*y1 + 40*y2 + 80*y3 + 40*y4 + 20*y5 + 30*y6 + 30*y7 + 40*y8 + 40*y10)-600;
eq20=x10*(50*y1 + 40*y2 + 40*y3 + 30*y4 + 50*y5 + 40*y6 + 20*y7 + 50*y8 + 30*y9)-600;
eqs=[eq1, eq2, eq3, eq4, eq5, eq6, eq7, eq8, eq9, eq10, eq11, eq12, eq13, eq14, eq15, eq16, eq17, eq18, eq19, eq20];
[x1,y1,x2,y2,x3,y3,x4,y4,x5,y5,x6,y6,x7,y7,x8,y8,x9,y9,x10,y10]=vpasolve(eqs,[x1,y1,x2,y2,x3,y3,x4,y4,x5,y5,x6,y6,x7,y7,x8,y8,x9,y9,x10,y10]);

请先登录，再回答此问题。

回答（1 个）

John D'Errico
John D'Errico 2021-12-23
Ran in:
You have 20 nonlinear equations in 20 unknowns. Do you seriously expect a symbolic solution? And of course, there will very possibly be many equally valid solutions, at least if any exist at all.
Even a numerical solution, as vpasolve would try to produce is going to take a significant amount of time. So why would you try to formulate this as s symbolic problem at all? If this has a hope of any solution in a reasonable amount of time, it would be using a numerical solver like fsolve. Could you try this?
opts = optimset('fsolve');
opts.MaxIterations = 10000;
xy = fsolve(@myfun,(2+rand(1,20)),opts)
No solution found. fsolve stopped because the relative size of the current step is less than the value of the step size tolerance squared, but the vector of function values is not near zero as measured by the value of the function tolerance.
xy = 1×20
1.0875 1.4248 1.3529 1.2332 0.9507 1.2385 1.0712 1.1458 1.3088 1.3934 1.0577 1.2665 1.1864 1.1060 1.3339 1.2795 1.1935 1.2017 1.2220 1.4634
myfun(xy)
ans = 20×1
9.8129 9.2448 13.0070 9.8242 7.7670 10.3430 10.1185 11.3644 10.6585 11.0053
I'll try it a second time, with a different random start point.
xy = fsolve(@myfun,(2+rand(1,20)),opts)
No solution found. fsolve stopped because the last step was ineffective. However, the vector of function values is not near zero, as measured by the value of the function tolerance.
xy = 1×20
0.8707 1.1409 1.0833 0.9875 0.7612 0.9917 0.8577 0.9174 1.0479 1.1157 1.3210 1.5817 1.4817 1.3812 1.6658 1.5980 1.4906 1.5008 1.5262 1.8276
myfun(xy)
ans = 20×1
9.8129 9.2448 13.0070 9.8242 7.7670 10.3430 10.1185 11.3644 10.6585 11.0053
So no solution was found. That does not mean a solution does not exist. But it does tell me what I expect, that there will be many multiple local stable points for the solver, and that if you will hope to find a solution, that would require a great deal of luck to have chosen good starting values.
But worse, see that despite trying different starting values, AND arriving at clearly differnt points for xy, in both cases, the values for myfun were close to identical. Does that tell me anything? YES> That suggests there may be no exact solution at all.
Would vpasolve do any better? Surely not, as you are asking this question in the first place. vpasolve requires a starting point. And the solution will be dependent on the start point, just like fsolve. But worse, vpasolve will not roduce an answer if an exact one does not exist. At least fsolve returns an answer. And vpasolve will take one hell of a long time to converge for this large of a problem.
All of that means there is likely no exact solution possible, even one that is close to exact.
function Z = myfun(XY)
Z = [XY(11)*(30*XY(2) + 40*XY(3) + 50*XY(4) + 60*XY(5) + 70*XY(6) + 30*XY(7) + 40*XY(8) + 50*XY(9) + 60*XY(10))-550;
XY(12)*(50*XY(1) + 40*XY(3) + 50*XY(4) + 30*XY(5) + 50*XY(6) + 40*XY(7) + 50*XY(8) + 60*XY(9) + 30*XY(10))-600;
XY(13)*(50*XY(1) + 30*XY(2) + 60*XY(4) + 20*XY(5) + 50*XY(6) + 40*XY(7) + 50*XY(8) + 40*XY(9) + 20*XY(10))-500;
XY(14)*(50*XY(1) + 40*XY(2) + 80*XY(3) + 50*XY(5) + 60*XY(6) + 50*XY(7) + 30*XY(8) + 40*XY(9) + 50*XY(10))-600;
XY(15)*(30*XY(1) + 50*XY(2) + 50*XY(3) + 60*XY(4) + 60*XY(6) + 50*XY(7) + 30*XY(8) + 30*XY(9) + 60*XY(10))-700;
XY(16)*(40*XY(1) + 50*XY(2) + 50*XY(3) + 50*XY(4) + 40*XY(5) + 60*XY(7) + 40*XY(8) + 50*XY(9) + 70*XY(10))-700;
XY(17)*(40*XY(1) + 50*XY(2) + 40*XY(3) + 50*XY(4) + 30*XY(5) + 80*XY(6) + 40*XY(8) + 50*XY(9) + 30*XY(10))-600;
XY(18)*(30*XY(1) + 30*XY(2) + 40*XY(3) + 40*XY(4) + 20*XY(5) + 30*XY(6) + 90*XY(7) + 40*XY(9) + 30*XY(10))-500;
XY(19)*(50*XY(1) + 40*XY(2) + 80*XY(3) + 40*XY(4) + 20*XY(5) + 30*XY(6) + 30*XY(7) + 40*XY(8) + 40*XY(10))-550;
XY(20)*(50*XY(1) + 40*XY(2) + 40*XY(3) + 30*XY(4) + 50*XY(5) + 40*XY(6) + 20*XY(7) + 50*XY(8) + 30*XY(9))-600;
XY(1)*(30*XY(12) + 40*XY(13) + 50*XY(14) + 60*XY(15) + 70*XY(16) + 30*XY(17) + 40*XY(18) + 50*XY(19) + 60*XY(20))-600;
XY(2)*(50*XY(11) + 40*XY(13) + 50*XY(14) + 30*XY(15) + 50*XY(16) + 40*XY(17) + 50*XY(18) + 60*XY(19) + 30*XY(20))-700;
XY(3)*(50*XY(11) + 30*XY(12) + 60*XY(14) + 20*XY(15) + 50*XY(16) + 40*XY(17) + 50*XY(18) + 40*XY(19) + 20*XY(20))-600;
XY(4)*(50*XY(11) + 40*XY(12) + 80*XY(13) + 50*XY(15) + 60*XY(16) + 50*XY(17) + 30*XY(18) + 40*XY(19) + 50*XY(20))-700;
XY(5)*(30*XY(11) + 50*XY(12) + 50*XY(13) + 60*XY(14) + 60*XY(16) + 50*XY(17) + 30*XY(18) + 30*XY(19) + 60*XY(20))-500;
XY(6)*(40*XY(11) + 50*XY(12) + 50*XY(13) + 50*XY(14) + 40*XY(15) + 60*XY(17) + 40*XY(18) + 50*XY(19) + 70*XY(20))-700;
XY(7)*(40*XY(11) + 50*XY(12) + 40*XY(13) + 50*XY(14) + 30*XY(15) + 80*XY(16) + 40*XY(18) + 50*XY(19) + 30*XY(20))-550;
XY(8)*(30*XY(11) + 30*XY(12) + 40*XY(13) + 40*XY(14) + 20*XY(15) + 30*XY(16) + 90*XY(17) + 40*XY(19) + 30*XY(20))-500;
XY(9)*(50*XY(11) + 40*XY(12) + 80*XY(13) + 40*XY(14) + 20*XY(15) + 30*XY(16) + 30*XY(17) + 40*XY(18) + 40*XY(20))-600;
XY(10)*(50*XY(11) + 40*XY(12) + 40*XY(13) + 30*XY(14) + 50*XY(15) + 40*XY(16) + 20*XY(17) + 50*XY(18) + 30*XY(19))-600];
end

类别

在 Help Center 和 File Exchange 中查找有关 Numeric Solvers 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by