Download MathematicalFunctionsSequences.mw

Edgardo S. Cheb-Terrab
Physics, Differential Equations and Mathematical Functions, Maplesoft

       Calculation of RSCR mechanism as a  solution to underdetermined system of nonlinear equations.  

RSCR.mw 

https://vk.com/doc242471809_376439263
https://vk.com/doc242471809_408704758

RCCC mechanism
https://vk.com/doc242471809_375452868

Solving DAEs in Maple

As I had mentioned in many posts, Maple cannot solve nonlinear DAEs. As of today (Maple 2015), given a system of index 1 DAE dy/dt = f(y,z); 0 = g(y,z), Maple “extends” g(y,z) to get dz/dt = g1(y,z). So, any given index 1 DAE is converted to a system of ODEs dy/dt = f, dz/dt = g1 with the constraint g = 0, before it solves. This is true for all the solvers in Maple despite the wrong claims in the help files. It is also true for MEBDFI, the FORTRAN implementation of which actually solves index 2 DAEs directly. In addition, the initial condition for the algebraic variable has to be consistent. The problem with using fsolve is that one cannot specify tolerance. Often times one has to solve DAEs at lower tolerances (1e-3 or 1e-4), not 1e-15. In addition, one cannot use compile =true for high efficiency.

The approach in Maple fails for many DAEs of practical interest. In addition, this can give wrong answers. For example, dy/dt = z, y^2+z^2-1=0, with y(0)=0 and z(0)=1 has a meaningful solution only from 0 to Pi/2, Maple’s dsolve/numeric will convert this to dy/dt = z and dz/dt = -y and integrate in time for ever. This is wrong as at t = Pi/2, the system becomes index 2 DAE and there is more than 1 acceptable solution.

We just recently got our paper accepted that helps Maple's dsolve and many ODE solvers in other languages handle DAEs directly. The approach is rather simple, the index 1 DAE is perturbed as dy/dt = f.S ; -mu.diff(g,t) = g. The switch function is a tanh function which is zero till t = tinit (initialization time). Mu is chosen to be a small number. In addition, lhs of the perturbed equation is simplified using approximate initial guesses as Maple cannot handle non-constant Mass matrix. The paper linked below gives below more details.

http://depts.washington.edu/maple/pubs/U_Apprach_FULL_DRAFT.pdf  

Next, I discuss different examples (code attached).

Example 1: Simple DAE (one ODE and one AE), the proposed approach helps solve this system efficiently without knowing the exact initial condition.

Hint: The code is given with a semicolon at the end of the most of the statements for educational purposes for new users. Using semicolon after dsolve numeric in classic worksheet crashes the code (Maplesoft folks couldn’t fix this despite my request).

Example 2:

This is a nickel battery model (one ODE and one AE). This fails with many solvers unless exact IC is given. The proposed approach works well. In particular, stiff=true, implicit=true option is found to be very efficient. The code given in example 1 can be used to solve example 2 or any other DAEs by just entering ODEs, AEs, ICs in proper places.

Example 3:

This is a nonlinear implicit ODE (posted in Mapleprimes earlier by joha, (http://www.mapleprimes.com/questions/203096-Solving-Nonlinear-ODE#answer211682 ). This can be converted to index 1 DAE and solved using the proposed approach.

Example 4:

This example was posted earlier by me in Mapleprimes (http://www.mapleprimes.com/posts/149877-ODEs-PDEs-And-Special-Functions) . Don’t try to solve this in Maple using its dsolve numeric solver for N greater than 5 directly. The proposed approach can handle this well. Tuning the perturbation parameters and using compile =true will help in solving this system more efficiently.

Finally example 1 is solved for different perturbation parameters to show how one can arrive at convergence. Perhaps more sophisticated users and Maplesoft folks can enable this as automatically tuned parameters in dsolve/numeric.

Note:

This post should not be viewed as just being critical on Maple. I have wasted a lot of time assuming that Maple does what it claims in its help file. This post is to bring awareness about the difficulty in dealing with DAEs. It is perfectly fine to not have a DAE solver, but when literature or standard methods are cited/claimed, they should be implemented in proper form. I will forever remain a loyal Maple user as it has enabled me to learn different topics efficiently. Note that Maplesim can solve DAEs, but it is a pain to create a Maplesim model/project just for solving a DAE. Also, events is a pain with Maplesim. The reference to Mapleprimes links are missing in the article, but will be added before the final printing stage. The ability of Maple to find analytical Jacobian helps in developing many robust ODE and DAE solvers and I hope to post my own approaches that will solve more complicated systems.

I strongly encourage testing of the proposed approach and implementation for various educational/research purposes by various users. The proposed approach enables solving lithium-ion and other battery/electrochemical storage models accurately in a robust manner. A disclosure has been filed with the University of Washington to apply for a provisional patent for battery models and Battery Management System for transportation, storage and other applications because of the current commercial interest in this topic (for batteries). In particular, use of this single step avoids intialization issues/(no need to initialize separately) for parameter estimation, state estimation or optimal control of battery models.

Download SolvingDAEsinMaple.mws

I'd like to pay attention to Maple packages for symmetric functions, posets, root systems, and finite Coxeter groups by John Stembridge. That was estimated as well qualified in Mathoverflow.

@Markiyan Hirnyk   I try not to use this package, as I think the results are not reliable enough. Here is the example, where instead of the three real roots it finds only one, despite the hint to look for the three roots:

restart;

DirectSearch:-SolveEquations(100^x=x^100, AllSolutions, solutions=3);

There are many other examples, particularly in discrete optimization in which it returns false results. Here is one example (well-known to you).

I would like to pay attention to an article " Sums and Integrals: The Swiss analysis knife " by Bill Casselman, where the Euler-Maclaurin formula is discussed.  It should be noted all that matter is implemented in Maple through the commands bernoulli and eulermac. For example,

bernoulli(666);

eulermac(1/x, x);

,

eulermac(sin(x), x, 6);

BTW, as far as I know it, this boring stuff is substantially used in modern physics. The one is considered in

Ronald Graham, Donald E. Knuth, and Oren Patashnik, Concrete mathematics, Addison-Wesley, 1989.

The last chapter is concerned with the Euler-MacLaurin formula.


           

You are teaching linear algebra, or perhaps a differential equations course that contains a unit on first-order linear systems. You need to get across the notion of the eigenpair of a matrix, that is, eigenvalues and eigenvectors, what they mean, how are they found, for what are they useful.

Of course, Maple's Context Menu can, with a click or two of the mouse, return both eigenvalues and eigenvectors. But that does not satisfy the needs of the student: an answer has been given but nothing has been learned. So, of what use is Maple in this pedagogical task? How can Maple enhance the lessons devoted to finding and using eigenpairs of a matrix?

In this webinar I am going to demonstrate how Maple can be used to get across the concept of the eigenpair, to show its meaning, to relate this concept to the by-hand algorithms taught in textbooks.

Ah, but it's not enough just to do the calculations - they also have to be easy to implement so that the implementation does not cloud the pedagogic goal. So, an essential element of this webinar will be its Clickable format, free of the encumbrance of commands and their related syntax. I'll use a syntax-free paradigm to keep the technology as simple as possible while achieving the didactic goal.

Notes added on July 7, 2015:

• This live webinar has already been presented and you can view the recording here: http://www.maplesoft.com/webinars/recorded/featured.aspx?id=922
  
  
• For one reaction to a point made in the webinar, see this post.

A small piece of code for fun before the weekend, inspired by some recent posts:

http://www.johndcook.com/blog/2015/06/03/mystery-curve/

http://mathlesstraveled.com/2015/06/04/random-cyclic-curves-5/

http://www.walkingrandomly.com/?p=5677

cycler := proc(t, k, p, m, n, T) local expr, u, v;
  u := exp(k*I*t);
  expr := exp(I*t) * (1 - u/m + I*(u^(-p))/n);
  v := 1 + abs(1/n) + abs(1/m);
  plots:-complexplot( expr, t = 0 .. T, axes = none,
                      view = [-v .. v, -v .. v] );
end proc:

cycler(t, 5, 3, 2, 3, 2*Pi);

And that can be made into a small application (needs Maple 18 or 2015),

Explore( cycler(t, k, p, m, n, T),
         parameters = [ k = -10 .. 10, p = -10 .. 10,
                        m = -10.0 .. 10.0, n = -10.0 .. 10.0,
                        [ T = 0 .. 2*Pi, animate ] ],
         initialvalues = [ k = 5, p = 3, m = 2, n = 3, T = 2*Pi ],
         placement = left, animate = false, numframes = 100 );

cyclerapp.mw

The animation of parameter T from 0 to 2*Pi can also be fun if the view option is removed/disabled in the call to complexplot. (That could even be made a choice, by adding an additional argument for calling procedure cycler and an accompanying checkbox parameter on the exploration.)

acer

This post is related to the this thread

The recursive procedure  PosIntSolve  finds the number of non-negative or positive solutions of any linear Diophantine equation

a1*x1+a2*x2+ ... +aN*xN = n  with positive coefficients a1, a2, ... aN .  

Formal parameters: L is the list of coefficients of the left part, n  is the right part,  s (optional) is nonneg (by default) for nonnegint solutions and  pos  for positive solutions.

The basic ideas:

1) If you make shifts of the all unknowns by the formulas  x1'=x1-1,  x2'=x2-1, ... , xN'=xN-1  then  the number of positive solutions of the first equation equals the number of non-negative solutions of the second equation.

2) The recurrence formula (penultimate line of the procedure) can easily be proved by induction.

The code of the procedure:

restart;

PosIntSolve:=proc(L::list(posint), n::nonnegint, s::symbol:=nonneg)

local N, n0;

option remember;

if s=pos then n0:=n-`+`(op(L)) else n0:=n fi;

N:=nops(L);

if N=1 then if irem(n0,L[1])=0 then return 1 else return 0 fi; fi;

add(PosIntSolve(subsop(1=NULL,L),n0-k*L[1]), k=0..floor(n0/L[1]));

end proc:

Examples of use.

Finding of the all positive solutions of equation 30*a+75*b+110*c+85*d+255*e+160*f+15*g+12*h+120*i=8000:

st:=time():

PosIntSolve([30,75,110,85,255,160,15,12,120], 8000, pos);

time()-st;

                                       13971409380

                                             2.125

To test the procedure, solve (separately for non-negative and positive solutions) the simple equation  2*x1+7*x2+3*x3=2000  in two ways (by the  procedure and brute force method):

ts:=time():

PosIntSolve([2,7,3], 2000);

PosIntSolve([2,7,3], 2000, pos);

time()-ts;

                47905

                47334

                 0.281

ts:=time():

k:=0:

for x from 0 to 2000/2 do

for y from 0 to floor((2000-2*x)/7) do

for z from 0 to floor((2000-2*x-7*y)/3) do

if 2*x+7*y+3*z=2000 then k:=k+1 fi;

od: od: od:

k; 

k:=0:

for x from 1 to 2000/2 do

for y from 1 to floor((2000-2*x)/7) do

for z from 1 to floor((2000-2*x-7*y)/3) do

if 2*x+7*y+3*z=2000 then k:=k+1 fi;

od: od: od:

k;

time()-ts; 

                   47905

                   47334

                   50.063

Another example - the solution of the famous problem: how many ways can be exchanged $ 1 using the coins of smaller denomination.

PosIntSolve([1,5,10,25,50],100);

                        292

 Number-of-solutions.mw

 Edit.  The code has been slightly edited 

The question of colorbars for plots comes up now and then.

One particular theme involves creating an associated 2D plot as the colorbar, using the data within the given 3D plot. But the issue arises of how to easily display the pair together. I'll mention that Maple 2015.1 (the point-release update) provides quite a simple way to accomplish the display of a 3D plot and an associated 2D colorbar side-by-side.

I'm just going to use the example of the latest Question on this topic. There are two parts to here. The first part involves creating a suitable colorbar as a 2D plot, based on the data in the given 3D plot. The second part involves displaying both together.

Here's the 3D plot used for motivating example, for which in this initial experiment I'll apply shading by using the z-value for hue shading. There are a few ways to do that, and a more complete treatment of the first part below would be to detect the shading scheme and compute the colorbar appropriately.

Some of the code below requires update release Maple 2015.1 in order to work. The colors look much better in the actual Maple Standard GUI than they do in this mapleprimes post (rendered by MapleNet).

f := 1.7+1.3*r^2-7.9*r^4+16*r^6:

P := plot3d([r, theta, f],
            r=0..1, theta=0..2*Pi, coords=cylindrical,
            color=[f,1,1,colortype=HSV]):

P;

Now for the first part. I'll construct two variants, one for a vertical colorbar and one for a horizontal colorbar.

I'll use the minimal and maximal z-values in the data of 3D plot P. This is one of several aspects that needs to be handled according to what's actually inside the 3D plot's data structure.

zmin,zmax := [min,max](op([1,1],P)[..,..,3])[]:

verthuebar := plots:-densityplot(z, dummy=0..1, z=zmin..zmax, grid=[2,49],
                                 size=[90,260], colorstyle=HUE,
                                 style=surface, axes=frame, labels=[``,``],
                                 axis[1]=[tickmarks=[]]):

horizhuebar := plots:-densityplot(z, z=zmin..zmax, dummy=0..1, grid=[49,2],
                                  size=[300,90], colorstyle=HUE,
                                  style=surface, axes=frame, labels=[``,``],
                                  axis[2]=[tickmarks=[]]):

Now we can approach the second part, to display the 3D plot and its colorbar together.

An idea which is quite old is to use a GUI Table for this. Before Maple 2015 that could be done by calling plots:-display on an Array containing the plots. But that involves manual interation to fix it up to look nice: the Table occupies the full width of the worksheet's window and must be resized with the mouse cursor, the relative widths of the columns are not right and must be manually adjusted, the Table borders are all visible and (if unwanted) must be hidden using context-menu changes on the Table, etc. And also the rendering of 2D plots in the display of the generated _PLOTARRAY doesn't respect the size option if applied when creating 2D plots.

I initially thought that I'd have to use several of the commands for programmatic content generation (new in Maple 2015) to build the GUI Table. But in the point-release Maple 2015.1 the size option on 2D plots is respected when using the DocumentTools:-Tabulate command (also new to Maple 2015). So the insertion and display is now possible with that single command.

DocumentTools:-Tabulate([P,verthuebar],
                        exterior=none, interior=none,
                        weights=[100,25], widthmode=pixels, width=420);

DocumentTools:-Tabulate([[P],[horizhuebar]],
                        exterior=none, interior=none);

We may also with to restrict the view, vertically, and have the colorbar match the shades that are displayed.

DocumentTools:-Tabulate([plots:-display(P,view=2..5),
                         plots:-display(verthuebar,view=2..5)],
                        exterior=none, interior=none,
                        weights=[100,25], widthmode=pixels, width=420);

DocumentTools:-Tabulate([[plots:-display(P,view=2..5)],
                         [plots:-display(horizhuebar,view=[2..5,default])]],
                        exterior=none, interior=none);

I'd like to wrap both parts into either one or two procedures, and hopefully I'd find time to do that and post here as a followup comment.

Apart from considerations such as handling various kinds of 3D plot data (GRID vs MESH, etc, in the data structure) there are also the matters of detecting a view (VIEW) if specified when creating the 3D plot, handling custom shading schemes, and so on. And then there's the matter of what to do when multiple 3D surfaces (plots) have been merged together.

It's also possible that the construction of the 2D plot colorbar is the only part which requires decent programming as a procedure with special options. The Tabulate command already offers options to specify the placement, column weighting, Table borders, etc. Perhaps it's not necessary to roll both parts into a single command.

3dhuecolorbarA.mw

acer

This post is related to the question. There were  proposed two ways of finding the volume of the cutted part of a sphere in the form of a wedge.  Here the procedure is presented that shows the rest of the sphere. Parameters procedure: R - radius of the sphere, H1 - the distance the first cutting plane to the plane  xOy,  H2 -  the distance the second cutting plane to the plane  zOy. Necessary conditions:  R>0,  H1>=0,  H2>=0,  H1^2+H2^2<R^2 . For clarity, different surfaces are painted in different colors.

restart;

Pic := proc (R::positive, H1::nonnegative, H2::nonnegative)

local A, B, C, E, F;

if R^2 <= H1^2+H2^2 then error "Should be H1^(2)+H2^(2)<R^(2)" end if;

A := plot3d([R*sin(theta)*cos(phi), R*sin(theta)*sin(phi), R*cos(theta)], phi = arctan(sqrt(-H1^2-H2^2+R^2), H2) .. 2*Pi-arctan(sqrt(-H1^2-H2^2+R^2), H2), theta = 0 .. Pi, color = green);

B := plot3d([R*sin(theta)*cos(phi), R*sin(theta)*sin(phi), R*cos(theta)], phi = -arctan(sqrt(-H1^2-H2^2+R^2), H2) .. arctan(sqrt(-H1^2-H2^2+R^2), H2), theta = 0 .. arccos(sqrt(R^2-H2^2-H2^2*tan(phi)^2)/R), color = green);

C := plot3d([R*sin(theta)*cos(phi), R*sin(theta)*sin(phi), R*cos(theta)], phi = -arctan(sqrt(-H1^2-H2^2+R^2), H2) .. arctan(sqrt(-H1^2-H2^2+R^2), H2), theta = arccos(H1/R) .. Pi, color = green);

E := plot3d([r*cos(phi), r*sin(phi), H1], phi = -arccos(H2/sqrt(R^2-H1^2)) .. arccos(H2/sqrt(R^2-H1^2)), r = H2/cos(phi) .. sqrt(R^2-H1^2), color = blue);

F := plot3d([H2, r*cos(phi), r*sin(phi)], phi = arccos(sqrt(-H1^2-H2^2+R^2)/sqrt(R^2-H2^2)) .. Pi-arccos(sqrt(-H1^2-H2^2+R^2)/sqrt(R^2-H2^2)), r = H1/sin(phi) .. sqrt(R^2-H2^2), color = gold);

plots[display](A, B, C, E, F, axes = none, view = [-1.5 .. 1.5, -1.5 .. 1.5, -1.5 .. 1.5], scaling = constrained, lightmodel = light4, orientation = [60, 80]);

end proc:

Example of use:

Pic(1,  0.5,  0.3);

In case anyone is interested, we recently posted a new application on the Application Center,

Time Series Analysis: Forecasting Average Global Temperatures

While interesting in itself (well, I think so, anyway), this application also provides tips and techniques for analyzing time series data in Maple, and shows how to access online data sets through the new data sets functionality in Maple 2015.

eithne

The trailers for the new Star Wars movie (Star Wars: The Force Awakens) introduced a new Droid called BB-8. This curious little guy features a spherical body and a controlled instrumented head. More recently, the BB-8 droid was showcased in a Star Wars celebration event and to many peoples' surprise it is real and not a CGI effect!

We have a Sphero robot from Orbotix here at the office, and there was an immediate connection between BB-8 and the Sphero. All that remains is to add the head!

Many have already put together their version of the BB-8, but I wanted to have a physical model that I can play with in a virtual environment and explore some design options.

Preparation:

To build a model of BB-8 like robotic system in MapleSim (Maplesoft's physical modeling software environment), I first needed a couple things in place before going forward:

1. A few simple CAD shapes (half-sphere, wheels)
   
   
2. A component to represent the contact between two spheres (both outside contact and inside contact)

I used Maple’s plottools package to build the CAD files I needed. First a half-spherical shape:

Then a wheel:

The next step was to create the contact component in MapleSim. I used a Modelica custom component to bring together vector calculations of normal and tangential forces with a variety of options for convenience into one component:

Build the model:

We start with a spherical shape contacting the ground:

Then we add two wheels inside it, and a hanging mass to keep the reference axis vertical when the wheels turn:

Learning from published diagrams showing the internal mechanism of a Sphero, another set of free wheels improves the overall stability when motion commands are given to the two active wheels:

Now this model can be used to move around the surface by giving speed commands to the individual motors that drive to the two bottom wheels. What is needed next is the head and the mechanism to move it around.

Since the head can move almost freely, independent of body rotation, it has to be controlled via magnetic contacts and a controlled arm.

First, we add the control arm:

Now we need to build the head.

The head has an identical triangle to the one at the end of the control arm. At each vertex there is a ball bearing that would slide on the surface of the main spherical body without friction. The magnetic force between the corresponding vertices of the two triangles is modeled via the available point-to-point force element in MapleSim.

Once assembled, the MapleSim model diagram looks like this:

...and our BB-8 droid looks like this:

Seeing the BB-8 in action:

Now that we have constructed our droid in MapleSim, we can animate and see it in action!

PLATFORM_2.mw

Two platforms (tripod) and abstract mechanism.

https://vk.com/doc242471809_379935473
https://vk.com/doc242471809_380837279
https://vk.com/doc242471809_378352452
https://vk.com/doc242471809_380217387

And six degrees of freedom of a rigid body



https://vk.com/doc242471809_383336706

Download Discovery_with_Maple.mw