Science topic

# Numerics - Science topic

Explore the latest questions and answers in Numerics, and find Numerics experts.

Questions related to Numerics

Hi folks!

Let's say that I have two lists / vectors "t_list" and "y_list" representing the relationship y(t). I also have numerically computed dy/dt and stored it into "dy_dt_list".

The problem is that "dy_dt_list" contains a lot of fluctuations, and that I know that it MONOTONOUSLY DECREASES out of a physical theory.

1) Is there is a simple way in R or Python to carry out a spline regression that reproduces the numerical values of dy/dt(t) in "dy_dt_list" as best it can UNDER THE CONSTRAINT that it keeps decreasing? I thus want to get a monotonously decreasing (dy/dt)_spline as the output.

2) Is there is a simple way in R or Python to carry out a spline regression that reproduces the numerical values of y(t) as best it can UNDER THE CONSTRAINT that (dy/dt)spline keeps decreasing? I thus want to get y_spline as the output, given that the above constraint is fulfilled.

I'd like to avoid having to reinvent the wheel!

P.S: I added an example to clarify things!

It will be very helpfull if one can provide me for example an excel sheet or like this showing how applying the NLDFT to calculate pore size distribution from experimental N2 adsorption isotherms ?

Regards,

I have a system dynamic that is non-smooth because it has several signum and absolute value functions in it (three-tank level control).

I can obviously choose different sigmoid functions to approximate the dynamics in a way that they become n-times continuously differentiable.

Can somebody point me out to a systematic approach for doing this?

I can think of many ways to approximate the functions but it is difficult for me to decide on the "best" one.

Right now I would base my decision on the following criteria:

- smoothness (differentiability class)

- numerical computation effort (especially concerning integration)

An example term that I would want to be smooth would be:

$$

sign(x_1 - x_2) \sqrt{abs(x_1-x_2)}

$$

The context of my question is nonlinear model predictive control.

How can I get the numerical solution of the Reynolds equation for the contact pin on disk pitted?

Good evening, i am studing a crack under a mode I (opening) loading located in a structure with non-homogeneous bondes, this problem can be reduced to a singular integral equation with a simple Cauchy type singular part and another complex-valued part, the obtained form is shown in the fingure below, i need to now what is the normaly used numerical procedure to solve this kind of SIEs that involve a complex Kernel part. thanks

Why my NUMERICAL EXPERIMENTS results almost no change with the increase of the proportion of labeled data?

data: orl-face-data. 50% of per class of the data is train_set,the other is test_set

A perennial question from my students is whether or not they should normalize (say, 0 to 1) a numerical target variable and/or the selected explanatory variables when using artificial neural networks. There seems to be two camps: those that say yes, and those that say it is unnecessary. But what is your opinion ....and importantly, why?

Is a PhD research work and Numerical FEA will be used. Additional sources of literature will be appreciated.

Hello

I use newmark method for solve numerical dynamic of equation in SDF system.

when solve liquid by MFS method and get interaction force,

This Force don't converge.

I would be grateful if answer this question.

sincerely

it seems like that in research gate. All added manuscript just labelled authorship numerically as 1st, 2nd. etc. However, some articles are actually co-first authorship. How to indicate that?

In a questionnaire questions are posed that give answers Yes / No, Likert scale and numerical measures. I need through all the answers, to obtain a single measure that quantifies the values in a unified way.

I have a curve and some mathematical model functions. I want to test each of these models, using TableCurve 2d, but I don't know how to insert the functions list of TableCurve.

For example, I want to add this function : Y=exp(-aX)

After converting my datas to .arff format,I have applied numeric to binary filter,but still i cant able to enable FPGrowth. For FPGrowth all the datas has to be converted to boolean values,for this i cant able to locate ignoreclass set to true,please help me to resolve this.

Barlat 1989 is a yield criterion for anisotropic material. To use this model, its constitutive parameters must be determined experimentally for specific material. In this model there is a parameter p which must determined numerically. But I do not know the procedure. I will appreciate if anybody know the procedure and let me know.

Iam carrying out research work in mines...I had solid data of mines with numerical..I want to reduce some factors from the data available..getting confused which approach is best suit to reduce the small factors rather than factor analysis/DOE. can ANNOVA can be used as factor reduction technique?

I am working on a 200-year numerical macro simulation. There are 2 treatment variables which means I have 4 different simulations. I want to find the effect of one treatment variable on the effect of the other. I am implementing a policy (one of the treatment variables), and I want to see the effect that the other treatment variable has on the effect of that policy. Is there a way to find this by differencing the time paths of the four treatments?

My two treatment variables are demographic change [D] (i.e. whether or not the labour force is affected by a baby boom-bust) and a policy [P].

Is it meaningful to look at an interaction term?

Policy effect: (ND|P) - (ND|NP)

Demographic effect: (D|NP) - (ND|NP)

Interaction effect: (D|P) - Policy Effect - Demographic Effect

My other idea was to look at the difference between the policy effect with demographic change and without it:

[(D|P) - (D|NP)] - [(ND|P) - (ND|NP)]

Is any of this meaningful?

Any help would be greatly appreciated.

Best,

Mackenzie

When solving PDEs numerically by discretization, we often require a test case. In most equations I encountered, the test cases already exist. But I need to know how I can define my test case easily. Moreover, must I require initial/boundary conditions to create a test case?

To convert fingerprint to numeric which can then be used for generating cryptographic key

You should consider for numerical solutions of nonlinear BVP.

I tried to solve it using finite difference method. But I am not able to obtain the solution.

Any work on "How we calculate uncertainty in numerical data like lattice constant and urbac energy"?

_{I am working on numerical wavelet analysis, recently for approximating some highly oscillating function in a very small support in finite domain some difficulty arise.}

RLS algorithm is accused of numerical instability due to the way Riccati update is formulated. One way to over come this issue is to use QR-decomposition of the auto-correlation matrix. However, it was discovered that if the symmetry of covariance matrix is maintained the RLS algorithms remains stable. Is it true? Can anyone provide a reference.

I understand Fisher score is a feature selection methods. Are there other methods similar to Fisher score for mostly numeric variables and 2 class problems.

How to get numerical solution to the poisson equation of Independent double gate MOSFET in COMSOL multiphysics ?

PFA for the equation and boundary conditions.

I have a scanned curve and I want a software that converts an image file showing a graph into numbers.

Could you suggest one?

dx/dt = A*x+B(u); y = C*x+D*u;

x(t0) = x0;

objective function J = int_{t=t0}^{tf} (1 + a1*u^2)dt

ending time tf is free.

constrain on the ending value of the first state: x_1(tf) = x_{1, tf}

where

A number of parameters in A, B, C, D vary linear-piecewisely with the first state x_1;

B(u) is nonlinear function of input u, e.g., B(u) = [u, u, u^2, u, u]^T

constrains:

input constrain: u_min < u < u_max;

output constrain: y<y_max;

i need to know which method is selected when simulating in fluent.

I want to do clustering and evaluating the clusters.I have mixture of nominal and discrete numerical data. I did clustering and now i want to use some useful indices to choose the best number of clusters. which of the internal indices i should use? I mention again that the data are mixture of nominal and discrete numerical data. Thanks

Can somebody give me an example to design a controller using h2 optimal control?

Consider an N-dim to N-dim iteration scheme:

x

^{n+1}= F (x^{n}) where x's are N-dim vectors and F is an N-dim -> N-dim polynomial-type function.From the physical background, we know that this iteration sequence should converge to a fixed point or a limit cycle. But somehow numerically it diverges. I suspect it's the accumulation of numerically errors. Is there a standard way to stablise the iteration process? Thanks!

I am a member a an fsae team and require to design a muffler to comply with the rules kindly help me in designing my muffler for a royal enfield 499cc engine.

I am currently during research on spam review detection. I have it based on supervised method SVM. I have to implement it as my research project. I want to know how I can convert a review text into numerical value to give as input to the SVM method. What features can I use for better performance of spam detection? Please give me your helpful suggestions.

Thank you in advance for your support.

Nowadays all of the major Fortran related numerical calculus have exactly mapped equivalent libraries in more modern language framework like Numerical Python (NumPy) and Scientific Python (SciPy).

What keeps physicists stuck with Fortran?

Performance?

Portability?

Scientific evidence?

I need to solve linear equation like this:

A*X=B

where A is square and singular.

Any one feasible solution is OK and just what I need.

It is solvable manually but too troublesome if the dimension is high.

So how can I solve such a given equation by tools like MATLAB?

I want to change the numeric attribute value for "age" to categories "young and old" by setting a cut off of 50. How to do it in weka?

How can I validate my experimental results of DMA. Is there any methods which I can follow. If there is any way then please send me the links

The numerical equations to calculate the drain fuel back from injector to diesel tank.

I'd like to know if anyone has any books/links /article on "solving boussinesq equations using numerical methods". Particularly some stuff on "MOL( Method of Lines)".

Thanks in advance.

Please suggest a simple approximation to TANH() function.

Take all the naturals and for each pair (i, j) with i<j create a directed edge i -> j with fixed probability p, thus obtaining an infinite DAG (directed acyclic graph) G. Let T be the transitive reduction of G (which is unique). How does the length of the shortest path *in T* from node h to node k (if any) grow with their 'numeric distance' (k - h)?

(The *longest* path from h to k should grow linearly with their numeric distance.)

I try to solve numerically the differential equation with mixed derivative. I found one possibility for the mixed derivative like y(i+1,j-1)-y(i,j-1)-y(i-1,j+1)+y(i-1,j). But may be there are another ways...

I have a set of both experimental and numerical values (say for every second). What are the different ways of finding the cumulative deviation between the values? Which can be considered the best among all?