Welcome Guest! To enable all features please Login. New Registrations are disabled.

Notification

Icon
Error

Login


Options
Go to last post Go to first unread
Offline OchkovVF  
#1 Posted : 02 October 2023 13:00:31(UTC)
OchkovVF


Rank: Advanced Member

Groups: Registered, Advanced Member
Joined: 06/04/2023(UTC)
Posts: 338
Man
Russian Federation

Was thanked: 26 time(s) in 25 post(s)
Do you know what method uses the roots function?
Look at the areas of attraction to the roots. White color - no solution.

Click to enlarge

Edited by moderator 02 October 2023 15:12:38(UTC)  | Reason: Not specified

Wanna join the discussion?! Login to your SMath Studio Forum forum account. New Registrations are disabled.

Offline Jean Giraud  
#2 Posted : 02 October 2023 17:09:09(UTC)
Jean Giraud

Rank: Guest

Groups: Registered
Joined: 04/07/2015(UTC)
Posts: 6,866
Canada

Was thanked: 981 time(s) in 809 post(s)
Originally Posted by: OchkovVF Go to Quoted Post
Do you know what method uses the roots function ?

Smath roots plugin is essentially Newton.
Descartes used it about a century before.
roots(f(x),x,init)= very accurate
thanks 1 user thanked Jean Giraud for this useful post.
on 03/10/2023(UTC)
Offline OchkovVF  
#3 Posted : 03 October 2023 09:09:08(UTC)
OchkovVF


Rank: Advanced Member

Groups: Registered, Advanced Member
Joined: 06/04/2023(UTC)
Posts: 338
Man
Russian Federation

Was thanked: 26 time(s) in 25 post(s)
Originally Posted by: Jean Giraud Go to Quoted Post
Originally Posted by: OchkovVF Go to Quoted Post
Do you know what method uses the roots function ?

Smath roots plugin is essentially Newton.
Descartes used it about a century before.
roots(f(x),x,init)= very accurate


I think the built-in roots function uses a different method than Newton. Which?
Offline Jean Giraud  
#4 Posted : 03 October 2023 15:03:09(UTC)
Jean Giraud

Rank: Guest

Groups: Registered
Joined: 04/07/2015(UTC)
Posts: 6,866
Canada

Was thanked: 981 time(s) in 809 post(s)
Originally Posted by: OchkovVF Go to Quoted Post
I think the built-in roots function uses a different method than Newton. Which ?

I get same 15 decimals from either Newton/roots.
Occasionally, roots fails not being able to initialize.


Offline grelkin2  
#5 Posted : 12 December 2023 13:26:01(UTC)
grelkin2


Rank: Advanced Member

Groups: Registered
Joined: 12/12/2020(UTC)
Posts: 35

Was thanked: 3 time(s) in 3 post(s)
polv1i.png
Рассмотрел данную систему в зависимости от задаваемой точности. Красным цветом обозначил точки, которые выходят в NaN, зеленным цветом точки, которые дают хоть одно решение. Часть решений показал желтыми линиями. При повышении точности "белые" области увеличиваются. Это те области, в которых при заданном числе шагов вычислений, не достигается никакого результата.
[ENG]
I considered this system depending on the specified accuracy. In red, I marked the points that go to NaN, in green, the points that give at least one solution. Some of the solutions were shown with yellow lines. As the accuracy increases, the "white" areas increase. These are the areas in which, with a given number of computational steps, no result is achieved.
[ENG]
Анимация polv1i.mov (1,891kb) downloaded 4 time(s).
Offline OchkovVF  
#6 Posted : 12 December 2023 20:58:01(UTC)
OchkovVF


Rank: Advanced Member

Groups: Registered, Advanced Member
Joined: 06/04/2023(UTC)
Posts: 338
Man
Russian Federation

Was thanked: 26 time(s) in 25 post(s)
Offline grelkin2  
#7 Posted : 13 December 2023 22:36:50(UTC)
grelkin2


Rank: Advanced Member

Groups: Registered
Joined: 12/12/2020(UTC)
Posts: 35

Was thanked: 3 time(s) in 3 post(s)
Originally Posted by: OchkovVF Go to Quoted Post

Спасибо за статью.
Если мы говорим о алгебраическом многчлене f(z)=aₙ*zⁿ+..+a₁*z+a₀=0, aᵢ=const i=0..n
zᵢ=bᵢ+i*cᵢ - i-ый корень уравнения, (bᵢ, cᵢ) искомые вещественные и мнимые части i-корня.
При подстановке и последующей группировки решаем систему из 2-х уравнений Re(f(z))=0 и Im(f(z))=0

Я не вникал глубоко в решение систем с большим n. Мне кажется при нахождение корня с приемлемой точностью,
проверяется его кратность, является корнем его сопряжение, и затем исходный многочлен делится на решение,
степень опускается и снова ищутся корни многочлена меньшей степени.

Для уравнения z^3-5*z^2-1.5=0, после замены z=a+b*i
-3*a*b^2-5*(a^2-b^2)+a^3-1.5=0
-b^3+3*a^2*b-10*a*b=0
Я использую метод Драгилева для решения нелинейных уравнений. Добавлем новую переменную n и константы,
которые будут зависеть от начального положения, то есть уравнение будет
f1(z,n)=z^3-5*z^2-1.5+(a0+a1*i)*n=0, n=1, a0 и a1 такие что при z0=a0+b0*i f1(z0,1)=0
-3*a*b^2-5*(a^2-b^2)+a^3-1.5+a0*n=0
-b^3+3*a^2*b-10*a*b+a1*n=0

Что касается полюсов, областей устойчивости, все это сложная вещь, связанная с правой частью ду
Для уравнений (я ввел дополнительный параметр n)
(x^2+y^2)^2-2*(x^2-y^2)+a0*n-3=0
((x-.5)^2+(y-.5)^2)^2-8*((x-.5)^2-(y-.5)^2)+a1*n=0
Система ду будет следующая
dx/dt = ( -2.0e+1*a0*y+-4.0*a0*(x*x)*y+-4.0*a0*(y*y*y)+9.0e+0*a0+4.0*(x*x)*a1*y+6.0e+0*a0*(y*y)+4.0*a1*(y*y*y)+-2.0*a0*x+2.0e+0*a0*(x*x)+4.0*a1*y+4.0e+0*a0*x*y);
dy/dt = ( 2.0*a0*y+-4.0*(x*x*x)*a1+7.0e+0*a0+-2.0e+0*a0*(y*y)+4.0*x*a1+4.0*a0*x*(y*y)+-1.2e+1*a0*x+-6.0e+0*a0*(x*x)+-4.0*x*a1*(y*y)+4.0*a0*(x*x*x)+-4.0e+0*a0*x*y);
dn/dt = ( 4.0e+0*x*(y*y)+-2.0e+1*(y*y*y)+-8.0*(y*y)+9.6e+1*x*(y*y*y)+8.0e+0*((x*x)*(x*x))*y+-8.0e+0*(x*(x*x)*(x*x))+8.0*((x*x)*(x*x))+-3.2e+1*x*y+-8.0*(x*x)+3.6e+1*x+1.2e+1*(x*x)*y+-2.8e+1*(x*x*x)+-1.6e+1*(x*x*x)*(y*y)+1.6e+1*(x*x)*(y*y*y)+9.6e+1*(x*x*x)*y+-8.0e+0*x*((y*y)*(y*y))+-2.8e+1*y+8.0e+0*(y*(y*y)*(y*y))+-8.0*((y*y)*(y*y)));



Users browsing this topic
Guest
Similar Topics
How to do root function in Smath studio (Questions)
by Aung 11/12/2018 13:34:19(UTC)
Unable to assign to variable in SS 0.98.5981 when using FindRoot Function (Bugs & Problems)
by ndtma 18/05/2016 15:20:15(UTC)
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.