Thanks to @salim-barzani, I became interested in the Laplace Adomian Decompositon method to solve partial differential equations. It produces a sum of analytical components that converge to the exact solution. Probably it is not that efficient for numerical solutions, but the possibility of finding an exact solution to nonlinear PDEs by finding a formula for the components and therefore the infinite sum is interesting.

The version here covers many of the common forms of PDEs, but is still a work in progress. The method for finding the Adomian polynomials could be improved by using one of the reccurence formulas in the literature. Extension to more PDE forms, ODEs etc are possible and I will attempt some of these before uploading an improved version to the application centre. In the meantime, feedback about bugs, usability etc. are appreciated.

Download Laplace_Adomian_Decomposition_procedure8b.mw

The link below goes to the Proceedings of the Maple 2024 Conference, which includes several articles that will be of interest to the readers of Maple Primes.

There may be one more paper coming in to the proceedings later as per policy; since most things are ready, away we go!

Proceedings of the Maple 2024 Conferenc

In Optical-Illusion--Impossible-Prisms, mcarvalho provides a Maple Learn worksheet to illustrate an intriguing optical illusion.  Here I do that illustration in Maple, and in 3D! 

The graphics below shows 4 bars.  Or is it 3 bars?

Download the worksheet, stare at the graphics for short while to grasp the nature of the optical illusion, and then rotate the 3D graph with the mouse and see the illusion fall apart.

optical-illusion.mw

Geometrical entertainment in the form of rolling without slipping, now inside a torus.
You can also print out the corresponding equations - these graphs are # in the text, but it is better to do this separately from the geometric animation, because textplot3d takes up a lot of resources.

Please consider it not as a Maple program, but simply as an idea for a corresponding algorithm.
for_TORUS_IN_TORUS_for.mw
 

The workbook includes the excel data file but I also included the worksheet and data file separate for versions that can't load workbooks.

Average_World_Temperature_workbook-.maple

Average_World_Temperature_since_1850_worksheet.mw

HadCRUT5.zip

Last week, DeepMind announced their new AlphaEvolve software along with a number of new results improving optimization problems in a number of areas. One of those results is a new, smaller, tensor decomposition for multiplying two 4 × 4 matrices in 48 scalar multiplications, improving on the previous best decomposition which needed 49 multiplications (Strassen 1969). While the DeepMind paper was mostly careful about stating this, when writing the introduction, the authors mistakenly/confusingly wrote:

For 56 years, designing an algorithm with rank less than 49 over any field with characteristic 0 was an open problem. AlphaEvolve is the first method to find a rank-48 algorithm to multiply two 4 × 4 complex-valued matrices

This gives the impression that this new result is the way to multiply 4 × 4 matrices over a field with the fewest number of field multiplications. However, it's not even the case that Strassen's 1969 paper gave a faster way to multiply 4 × 4 matrices. In fact, Winograd's 1968 paper on computing inner products faster could be applied to 4 × 4 matrix multiplication to get a formula using only 48 multiplications. Winograd uses a trick 

which changes two multiplications of ab's into one multiplication of ab's and two multiplications involving only a's or only b's. The latter multiplications can be pre-computed and saved when calculating all the innerproducts in a matrix multiplication

    # i=1..4, 4*2 = 8 multiplications
    p[i] := -A[i, 1]*A[i, 2] - A[i, 3]*A[i, 4];
    # 4*2 = 8 multiplications
    q[i] := -B[1, i]*B[2, i] - B[3, i]*B[4, i];

    # i=1..4, j=1..4, 4*4*2 = 32 multiplications
    C[i,j] = p[i] + q[j] 
            + (A[i,1]+B[2,j])*(A[i,2]+B[1,j])
            + (A[i,3]+B[4,j])*(A[i,4]+B[3,j])

It is simple to verify that C[i,j] = A[i,..] . B[..,j] and that the above formula calculates all of the entries of C=A.B with 8+8+32=48 multiplications.

So, if Winograd achieved a formula of 48 multiplications in 1968, why is the DeepMind result still interesting? Well, the relative shortcoming of Winograd's method is that it only works if the entries of A and B commute, while tensor decomposition formulas for matrix multiplication work even over noncommutative rings. That seems very esoteric, but everyone's favorite noncommutative ring is the ring of n × n matrices. So if a formula applies to a matrix of matrices, that means it gives a recursive formula for matrix multiplication - an tensor decomposition formulas do just that.

The original 1969 Strassen result gave a tensor decomposition of 2 × 2 matrix multiplication using only 7 scalar multiplications (rather than the 8 required by the inner product method) and leads to a recursive matrix multiplication algorithm using O(N^log[2](7)) scalar multiplications. Since then, it has been proved that 7 is the smallest rank tensor decompostion for the 2 × 2 case and so researchers have been interested in the 3x3 and larger cases. Alexandre Sedoglavic at the University of Lille catalogs the best tensor decompositions at https://fmm.univ-lille.fr/ with links to a Maple file with an evaluation formula for each.

The previous best 4 × 4 tensor decomposition was based on using the 2 × 2 Strassen decomposition recursively (2 × 2 matrix of 2 × 2 matrices) which led to 7 2 × 2 matrix multiplications each requiring 7 scalar multiplications, for a total of 49. The new DeepMind result reduces that to 48 scalar multiplications, which leads to a recursive algorithm using O(N^log[4](48)) scalar multiplications: O(N^2.7925) vs. O(N^2.8074) for Strassen. This is a theoretical improvment over Strassen but in 2024 the best known multiplication algorithm has much lower complexity: O(N^2.3713) (see https://arxiv.org/abs/2404.16349). Now, there might be some chance that the DeepMind result could be used in practical implementations, but its reduction in the number of multiplications comes at the cost of many more additions. Before doing any optimizations I counted 1264 additions in the DeepMind tensor, compared to 318 in the unoptimized 4×4 Strassen tensor (which can be optimized to 12*7+4*12=132). Finally, the DeepMind tensor decomposition involves 1/2 and sqrt(-1), so it cannot be used for fields of characteristic 2, or fields without sqrt(-1). Of course, the restriction on characteristic 2 is not a big deal, since the DeepMind team found a 4 × 4 tensor decomposition of rank 47 over GF2 in 2023 (see https://github.com/google-deepmind/alphatensor).

Now if one is only interested in 4 × 4 matrices over a commutative ring, the Winograd result isn't even the best possible. In 1970, Waksman https://ieeexplore.ieee.org/document/1671519 demonstrated an improvement of Winograd's method that used only 46 multiplications as long as the ring allowed division by 2. Waksman's method has since been improved by Rosowski to remove the divisions see https://arxiv.org/abs/1904.07683. Here's that nice compact straight-line program that computes C = A · B in 46 multiplications.

And, here attached are five Maple worksheets that create the explicit formulas for the 4 × 4 matrix methods mentioned in this post, and verify their operation count, and that they are correct.

Strassen_444.mw  Winograd.mw  DM_444.mw Waksman.mw Rosowski.mw

If you are interested in some of the 2022 results, I posted some code to my GitHub account for creating formulas from tensor decompositions, and verifying them on symbolic matrix multiplications: https://github.com/johnpmay/MapleSnippets/tree/main/FastMatrixMultiplication

The clear followup question to all of this is:  does Maple use any of these fast multiplication schemes?  And the answer to that is mostly hidden in low level BLAS code, but generally the answer is: No, the straightforward inner-product scheme for multiplying matrices optimized memory access thus usually ends up being the fastest choice in most cases.

For a long time, triggered by a disagreement with one of my teachers, I wanted to demonstrate that Euler equations are not absolutely necessary to reproduce gyroscopic effects. Back then, there were no computer tools like Maple or programming languages with powerful libraries like Python. Propper calculations by hand (combining Newton’s equations and vector calculus) would have required days without guarantee of immediate success. Overall, costs and risks were too high to go into an academic argument with someone in charge of grading students.

Some years ago, I remembered the unfinished discussion with my teacher and simulated with MapleSim the simple gyroscope with two point masses that I had in mind at the time. It took only 10 minutes to demonstrate that I was right. At least I thought so. As I discovered recently when investigating the intermediate axis theorem, MapleSim derives behind the scenes Euler equations. This devalued the demonstration.

This post is about a second and successful attempt of a demonstration with Maple employing Lagrangian mechanics.  A rotating system of three point masses connected by rigid struts is used. The animation below from the attached Maple worksheet exactly reproduces a simulation of an equivalent T-shaped structure of three identical masses presented here.

Lagrangian Mechanics

The worksheet uses Lagrangian mechanics to derive equations of motion. Only translational energy terms are used in the Lagrangian to prevent Euler equations from being derived. To account for the bound motion of the three point masses, geometric constraints with Lagrange multipliers were added to the Lagrangian L of the system. This lead to a modified Lagrangian L’ that can be used with dedicated Maple commands to derive with little effort a set of Lagrange’s equations of the first kind and the corresponding constraints

(source https://en.wikipedia.org/wiki/Lagrangian_mechanics)

For the system of three point masses the above equations lead to 9 coupled second-order ordinary differential equations (ODEs) and 3 algebraic constraints 

(Maple output from the command Physics,LagrangeEquations)

where x_i, y_i and z_i are the components of the position vectors _i of the masses 1 to 3 and the l_i,j are the constraints between the masses i and j, and b and h are the base and the height of the triangle.

The 12 equations together are also referred to as differential algebraic equations (DAEs). Maple has dedicated solvers for such systems which make implementation easy. The most difficult part is setting the initial conditions for all point masses. In this respect MapleSim is even easier to use since not all initial conditions have to be exactly defined by the user. MapleSim also detects constraints that allow for a simplification of the problem by reducing the number of variables to solve. This leads automatically to 3 instead of 12 equations to be solved. Computational effort is reduced in this way significantly.

Newtonian Mechanics

One could argue now that the above is a demonstration with Lagrangian mechanics and not with Newtonian mechanics. To treat the system in a Newtonian way, the masses must be isolated and internal forces acting on each mass via the struts are applied to each mass and effectively become external forces. This leads to 9 ODEs with 27 unknows

Following actio=reactio for each of the (massless) struts reduces the number of unknows to 18

To solve for the 18 unknows, 9 more relations are required. 3 algebraic constraints that keep the distance between the masses constant have already been listed in the previous section. 6 further algebraic constraints can be established from the fact that the force vectors point towards the opposing mass (see also below).
The effort to solve this system of equations will be even greater but with the benefit of having information about the internal forces in the system. Before making this effort, it is advisable to take a closer look at the equations of motion derived so far.

Forces and the "mysterious" Lagrange Multipliers

Rearranging equations of motion from Lagrangian mechanics to

and comparing this to the equations of motion from Newtonian mechanics yields in vector notation 

or more general for the forces

where the _i are the position vectors of the individual masses and the  are the constraint forces between them.  

In the case of a triangle formed by struts all internal forces must act in the direction of the edges of the triangle. If they would not act this way, opposing pairs of  and   would create a torque around the struts which would lead to an infinite angular acceleration of the massless struts. The above equation confirms this reasoning: The internal forces act in the direction of the difference vectors between the position vectors of the masses (which describe the edges) and scale with l_ij.

The beauty of the Lagrange multipliers in this case is that they hit 3 birds (three components of the vectors ) with one stone. This reduces the number of unknowns.

However, the Lagrange multipliers are somehow mysterious because they do not represent a physical quantity, but they can be used to calculate meaningful and correct physical results.

What makes them even more mysterious is the fact that positions constraints can be expressed in different ways. In the above example the square of the distance between the masses is kept constant instead of the distance. There are many more possibilities to formulate the constraint of constant distance and each of them results in different multipliers l_ij with different units. In principle they should all work equivalently but might not all be usable with dedicated solvers.

According to the above equation, the internal forces in a strut scale with the Lagrange multipliers and the length of the strut. During the back and forth flip of the triangle in the above animation the forces vary which can be appreciated from the l_ij in the following plot. 

Some observations:

• At the start, only the strut between the red and the green masses is tensed by centrifugal forces. This would have been intuitively expected.
• At the start, the broken symmetry in the initial conditions is already visible by the imbalance of the forces in the two other struts.
• At no time two forces are zero
• There is never compression in two struts at the same time. The existence of compression forces renders any attempt to replace the struts by cables useless
• Plotted together in 3D the Lagrange multipliers describe a seemingly perfect circle.

The last observation in particular shows that both the Lagrange multipliers and the IAT still hold some secrets that need to be clarified:

• Is it an exact circle? How to prove that?
• Does a circle appear for all initial conditions and geometries (obtuse isosceles triangles) when the intermediate axis theorem manifests?
• What does the circle represent? Is it a kind of an invariant between the Lagrange multipliers that allows the calculation of one force when the two others are known?
• Is there an analytical representation of the circle as a function of the geometry and the initial conditions?
• What determines the center, the radius and the orientation of the circle?

Conclusion

Overall, the approach of adding non-physical terms that are zero to a physical quantity (the difference between kinetic and potential energy) to derive something meaningful is not obvious at all and underlines the genius of Lagrange. For a system of three bound masses it could be used to calculate internal forces as opposed to the more common use of calculating external constraint forces. Beyond the lack of fully satisfying intuitive explanations, the IAT still offers unanswered scientific questions.

PS.: 

• The exercise was a nice cross-check of MapleSim and Maple.
• dsolve and odeplot are awsome

IAT_without_Euler_equations.mw

My maplet (author: Xavier Cormier) "sudokuvision_multithread3" allows you to generate, solve, and play nxm region sudoku puzzles. It is possible to import lists of puzzles in .csv, .sdm, and .sdk format. It loads a grid in line format. It saves puzzles in .txt and .csv format to import them on Android (Andoku3, OpenSudoku; you must re-save the file in .sdm) and on iOS (SmartSudoku, Good Sudoku). It saves game puzzles (in numbers or colors) and their solution as a .gif image.
Two possible interfaces.
Multithreading is used to generate sudoku puzzles in parallel on multiple threads: we use sequences for the random number and the number of sudokus for each thread:
e.g., [4521,7458]=s1 for the random number and [2,2]=s2 for the number of sudokus. We then generate 2[i] sudokus for the ith thread. with s1[i] as a random integer.
The maplet is compatible with Maple 2021 but not with Maple 2015.

Go to my Maple Cloud link :

https://maple.cloud/app/6283420531032064/sudoku+multithread3?key=8DBF1E6D71EF423B9BE883C27FF0FD892C14B4FCD25B4920A29AFE4E4B4363B8

My maplet (author :Xavier Cormier) studies the partitioning of a set and the decomposition of an integer into a sum.
It can be used in the case of a dice roll.

It determines:
- the number of times each face is obtained for n rolls for an m-sided die.
- the decomposition of an integer nx into the sum of mx integers.
- the partitioning of a set into u non-empty, disjoint pairwise parts.
- the partitions of an integer as well as the dice rolls having a given sum and the probability (equiprobability) of obtaining this sum.

Go to my link on Maple Cloud :

https://maple.cloud/app/6553582941372416/decomposition+d%27un+entier+en+somme+et+partage+d%27une+liste+en+parties+-+app+pour+lanc%C3%A9s+de+d%C3%A9s?key=8533B69313624ADCBE181600B15A1F4F5DDBBFAE86694B479CE492515459A2C9

Hello everyone,

I am creating this post to begin a thread where I will share a series of worksheets on important topics in Complex Analysis, written as part of my notes for my classes. Complex_Analysis_Notes.pdf

The planned sections include:

• Section 1: Infinite Series

• Section 2: Power Series

• Section 3: The Radius of Convergence of a Power Series

• Section 4: The Riemann Zeta Function and the Riemann Hypothesis

• Section 5: The Prime Number Theorem

Each worksheet will include calculations, plots, and examples using Maple to illustrate key ideas.

✅ I plan to upload one worksheet every week to keep a steady pace and allow time for discussion and feedback between posts.

I hope this thread will be helpful both for learning and for deeper exploration.
Feel free to comment, suggest improvements, or ask questions as I post the materials.

Thank you!

Download infinite_series.mw

Approximation_of_ODEs_with_Cubic_Splines.mw

Download New_in_Physics_2025.mw

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

Download Einstein_Equations_From_First_Principles.mw

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

I’m absolutely delighted to announce the launch of Maple 2025!

Although you see a new release every year, new features take anything from a few fast-paced weeks to develop, to months of careful cultivation.

Working on so many features in parallel, each with varying time scales, isn't easy! We have to fastidiously manage and track our work.

So it's easy to lose ourselves in the daily minutiae of software development. To help us maintain perspective, we constantly ask ourselves questions like:

• What user problem are we solving and how often does this problem occur?
• Can we validate our proposed solution with preliminary user feedback?
• Is this a solution to a problem that doesn't exist and will never exist, or are we pre-empting a future need?
• Are we offering value to our users?

Given the answers, we course-correct to make sure we stay on track for our central mission - to make you happy, and to keep you coming back year-after-year.

With Maple 2025, I think we've smashed that goal. We have many new features that'll appeal to many different types of users - from students, educators and mathematicians, to engineers, scientists and technical professionals

Let me walk you through some of my personal highlights.

It’ll be difficult for anyone to miss this - Maple 2025 has a new interface! It’s a ribbon-based UI that look clean and contemporary, and helps you find and discover tools more quickly than before.

You have large, meaningful icons.

Items are logically grouped.

The ribbons is contextual. If you click on a plot, you get new tabs for interacting with and drawing on the plot.

A new Education tab collects pedagogical resources that were scattered around the interface in prior releases.

This is the biggest visual overhaul to Maple in many years. We hope you like it! 

We also appreciate that changes in look and feel can be divisive. Please rest assured that we will refine and finesse the interface with each successive release; your comments and suggestions are most welcome.

The new interface is available on Windows and Linux, and as a technology preview on Mac.

The right arrow key on my keyboard is wearing out…and it’s all because of Maple. I’m knee deep in Maple nearly every day entering equations, and I’m always using right-arrow to move the cursor. It gets kind of tedious!

This anecdote reflects some investigative work we did. We comprehensively examined our internal library of thousands of Maple worksheets and discovered that these three input patterns are extremely common.

Previously, you’d use the right-arrow key to move the cursor out of the exponential, division or subscript.

Now, in Maple 2025, when you

• type ^, /, or enter a literal subscript with a double-underscore,
• followed by a number or symbol
• and then input another operator (such as +)

the operator is automatically inserted on the baseline (except when y = 1).

Of course, you can also make the cursor return to or stay in the exponent or denominator with a simple keystroke, when that is what is needed.

This is one of those little quality of life refinements that I’m very fond of - it’s a little visual and usability dopamine hit.

The sum command (and its typeset form) now indexes into vectors without you needing to spam unevaluation quotes all over your expression.

We’ve been integrating units deeper into the Maple system, release after release. Much of this is driven by our engineering users.

A few releases ago, we made int(numeric) compatible with units. With Maple 2025, you can now numerically differentiate  expressions and procedures that have units.

I’m a grizzled thermodynamics hack, so here’s an example in which I calculate the specific heat capacity of water by differentiating enthalpy with respect to temperature (and then confirm the result with the built-in value):

This is in addition to many other improvements to the units experience.

Although this is a part of Maple that I don’t touch often (my colleague Karishma takes point on the education side), I REALLY wish I’d had this when I was struggling with math.

You can now automatically generate unlimited variants of the same problem for students to solve with the Try Another feature, which has been added to Maple’s Check My Work tools (another feature I really could have used!). This is available for many common math principles, including factorization, simplification, integration and more.

This is just one of the improvements in Maple 2025 for teaching and learning.

 If you’ve ever found yourself going back and forth (and back and forth) between two large, almost identical-looking Maple expressions, trying to figure out how they are different, you’re going to love this one.  ExpressionTools is a new package that lets you compare the differences between two expressions.

I really like the use of color to highlight differences. Less squinting at the screen!

You can now run Maple Flow worksheets from Maple (you don’t need Maple Flow installed to do this). You can send parameters into the Flow worksheet and extract your desired results.

This means you can use the entire flexibility of Maple to analysis and manipulate your Flow worksheet. You could, for example:

• Attach a Flow worksheet to a Maple workbook and create an interactive application
• Carry out parameter studies of a Flow worksheet by evaluating it over many parameter sets in Maple
• Create an Excel interface for a Flow worksheet using the Maple add-in for Excel

Simplify is one of those functions that literally tens of thousands of people use each day. Every time we make an incremental improvement, the cumulative benefits across our entire user base are significant.

We’ve refined simplify in a number of critical ways. For example, simplify now recognizes when exponentials can be profitably converted to hyperbolic trig functions:

The analysis of many scientific phenomena result in Laplace transforms that do not have a symbolic inverse which can be expressed in terms of elementary functions. This includes applications in heat transfer, fluid mechanics, fractional diffusion processes, control systems and electrical transmission.

For example, this monster Laplace transform results from an analysis of voltage on a transmission line:

You can now numerically invert this transform courtesy of an enhancement to inttrans:-invlaplace - a fast quadrature method.

I’ve saved what I think has the most future potential for last.

I’m sure nearly all of you have experimented with the various AI tools. They’re an inevitable part of our present and future, whether we're comfortable with it or not.

This is something we've been mulling over for some time.

• In Maple 2019, the DeepLearning package made its debut. This package provides tools for machine learning, supporting operations such as classification and regression using neural networks.
• In Maple 2024, we introduced an AI-powered formula lookup feature.

In Maple 2025, we’re giving you an early-stage technology preview of AI-powered document generation.

You can automatically generate worksheet content by prompting an AI, and then gradually refine the content

If you’re an educator, you might want some content that describes applications of calculus. So you might ask the AI “How do I derive the formula for the area of a circle” by entering your prompt into this text box:

This is the worksheet content that may be returned:

If you’re structural engineer who wants to know how to calculate the hardness of concrete, you might ask the AI: “How do I calculate the compressive strength of slow hardening concrete as a function of time? Use the CEB-FIP Model Code 90. Include a worked example with Maple code”.

This worksheet content that could be generated (note the live Maple code):

We’re labelling AI-generated worksheet content as a technology preview. You might see

• text that might be misleading (but sounds plausible)
• code that doesn’t work (but looks plausible)
• or different results each time you click “Generate Document”

For the moment, I would not rely on AI-generated worksheet content without realistic expectations, a healthy dose of scepticism and a modicum of detached analysis. But AI models are rapidly growing in robustness, and we want to position ourselves to best exploit their future potential. The next few years will be VERY exciting.

We can never cover everything in a short blog post like this. So if you want to know more, head on over to the What’s New pages for Maple 2025!

Hi MaplePrimes,
I made a quick code to calculate Aliquot Sequences.

The code works.

I was inspired by a Numberphile YouTube video.

see

aliquot_sequence_cut_1.mw

aliquot_sequence_cut_1.pdf

This should be helpful.

Regards,

Matt