Differential Equation Solver for TypeScript

A simple Typescript Library for Approximating Solutions to Ordinary Differential Equations using Euler's method.

This project is very simple, but is a proving ground for my future ideas for Mathematics in JS/TS.

Installation

Installation is easy using NPM:

• Install it directly from NPM

    npm i differential-solver

• From GitHub

    npm i github.com/cooperw824/differential-solver

• From Source

    git clone github.com/cooperw824/differential-solver
  
    cd path/to/project
  
    npm i path/to/differential-solver

Usage

To start, create a new DifferentialEquation object:

new DifferentialEquation(
differential:(x: number, y:number) => number,
independentVarInitialCondition:number,
dependentVarInitialCondition:number )

The DifferentialEquation class takes three parameters:

• The Differential Equation to Evaluate (Required)

  • A function with two number inputs that outputs a number

  (x: number, y: number) => number;

• The Independent Variable's (x-value) Initial Condition (Required)

  • A number that indicates the initial value for X

• The Dependent Variable's (y-value) Initial Condition (Required)

  • A number that indicates the initial value for y

In typescript:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

In JavaScript:

import {DifferentialEquation} from 'differential-solver'; // ES Modules Import

// const { DifferentialEquation } = require('differential-solver'); // CommonJS Import

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

Approximating Solutions to Differential Equations

eulersMethod(
targetIndependentVariable: number,
steps: number,
options?: EulersMethodOptions)

To approximate the solution of Differential Equation at certain independent variable, make a call to the Euler's Method function

The Euler's Method function has two different modes: recursive and non recursive, they both return the same answer, but if your desired number of steps is too large you may want to use the non-recursive version.

• Parameters
  • targetIndependentVariable: number (Required)
    • The x-value to approximate a solution for
    • f(targetIndependentVariable) ≈ output
  • steps: number (Required)
  • options: EulersMethodOptions (Optional)
    • recursive: Boolean (Optional)
      • Should the recursive or non recursive version of the algorithm be used? Defaults to false.
    • rounding: Number (Optional)
      • Round to this many decimals. Defaults to 3.

Non-recursive version

By default the non-recursive function runs:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

// approximates f(1) in 3 steps, when dy/dx = x + y and f(0) = 1 
let approximation = differential.eulersMethod(1, 3)

console.log(approximation)
// prints: 2.741

To add more decimals after the decimal point:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

// approximates f(1) in 3 steps, when dy/dx = x + y and f(0) = 1 
let approximation = differential.eulersMethod(1, 3, {rounding: 10})

console.log(approximation)
// prints: 2.7407407407

Recursive Version

Using the recursive version is as easy as setting the recursive option to true:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

// approximates f(1) in 3 steps, when dy/dx = x + y and f(0) = 1 
let approximation = differential.eulersMethod(1, 3, {recursive: true})

console.log(approximation)
// prints: 2.741

To add more decimals after the decimal point:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

// approximates f(1) in 3 steps, when dy/dx = x + y and f(0) = 1 
let approximation = differential.eulersMethod(1, 3, {rounding: 10, recursive: true})

console.log(approximation)
// prints: 2.7407407407

Evaluating a Derivative at a Point

evaluateDifferential(independentVar?: number,
dependentVar?: number,
rounding?: number)

Returns the value of the derivative at the given point

• Parameters:
  • independentVar: number (Optional)
    • The x value to evaluate the derivative at. Defaults to the independent variable's initial condition
  • dependentVar: number (Optional)
    • The y value to evaluate the derivative at. Defaults to the dependent variable's initial condition
  • rounding: number (Optional)
    • How many numbers after the decimal point to round to. Defaults to not rounding.
• Returns
  • differential(independentVar, dependentVar)
  • The derivative at the given point x,y

Using initial condition:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

let derivative = differential.evaluateDifferential()

console.log(derivative)
// Prints 1

Using given values:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

let derivative = differential.evaluateDifferential(3.14159, 2.718)

console.log(derivative)
// Prints 5.85959

Using given values, and given rounding:

import {DifferentialEquation} from 'differential-solver';

let differential  = new DifferentialEquation((x + y) => x+y, 0, 1);

let derivative = differential.evaluateDifferential(3.14159, 2.718, 3)

console.log(derivative)
// Prints 5.86
// 5.85959 rounds to 5.86

Building From Source

A build is packaged with the code in the dist folder, but to build the program on windows you can just run

    npm run build

for other operating systems you may need to adjust the fixup script and copy command.

Contributing

This is a very small library, but acts as my proving ground for building future JS / TS modules. I have a few ideas for expanding this, but if you have any issues or ideas please open a GitHub issue.

If you have a feature you would to add, fork the repository, implement your feature, add / verify tests, and then open your pull request.