JCM_SS: Janus Cosmological Modele Simple Simulator

JCM_SS is Janus Cosmological Model N-Body simulation.

For more information regarding Janus model please see form the author(s):

• http://www.jp-petit.org
• Youtube video channel https://www.youtube.com/user/JPPETITofficiel
• http://www.jp-petit.org/science/Le_Modele_Cosmologique_Janus.pdf

Introduction of M- space objects

JCM_SS is a simple N-Body simulation (without collision management and me without .

As any N-Body simulation, we manage "M+" positive mass space objects. But in JCM_SS we add in the model a new type of space objects know as "M-" who have a "negative mass".

Interaction between space objects follow those rules:

• M+ interact with M+ following Newton gravity,
• M- interact with M- following Newton gravity,
• M+ interact with M- following a "reverse" Newton gravity (they repel).

Newton force is F=G * m * m' / r / r where :

• G is gravity constant,
• M is the mass of first objects
• m' is the mass of the other object
• r is the distance between particules

The newton function is in 1/r2 which lead for very close objects to an "infinite force". We add a "smooth parameter" which consider that two "to close objects" where r < tooClose join for new object with a mass of m + m' (except opposite mass which never join).

At the opposite, too "far away" object have almost no interaction force is "too close to nothing". We add another "smooth" parameter to r > tooFar which indicate to the system to not consider this interaction (we save some compute).

How-to

Compilation is something like:

make

To run the simulaton, first create a config_file.json and launch the program:

>./jcm_ss <config_file.json>

Configuration options

A config file is a simple json file:

{
  "Save Dir": "tmp\/",
  "Save File": "file.pov",
  "Universe Radius": 1000,
  "Nbr Tic": 1000,
  "Too Close": 1000,
  "Too Far": 1000,
  "Object Size": 1000,
  "Initial Space Objects": [
    {"x": -2, "y": 0, "vx": -1, "vy":0, "mass": 5},
    {"x": 2, "y": 0, "vx": 1, "vy":0, "mass": -5},
  ]
}

Where:

• Save Dir: is the directory where result json will be saved
• Save File: is the suffix of output file. Each tic, the program will create json file name <Save Dir>/<tic #>-<Save File> that will be able to be parsed by a renderer
• Universe Radius: is the radius of the Universe
• Nbr Tic: nbr of time the simulator will do the simulation
• Too Close: smoother see previous section
• Too Far: smoother see previous section
• Initial Space Objects: are the initial condition of your simulation

You can generate your configuration file using initCond.rb <num objects> and then ./jcm_ss <num objects>.json

Output

The output of the simulation will create several file in Save Dir. File prefix is the <tic number>- and suffis is user define by <Save File> parameter.

{
  "Tic": 0,
  "Universe Radius": 1000,
  "Space Object": [
    {"x": -2.000000, "y": 0.000000, "Size": 1000, "Sign": "M-"},
    {"x": 2.000000, "y": 0.000000, "Size": 1000, "Sign": "M+"},
  ]
}

Where Tic is the number of tic starting a 0. Universe Radius is the radius of the universe and each Space Object is an array of object, where x and y are the positon of the object in the referential, size is the size of the object (this is relative to its mass), and sign indicate if it's an M+ or M- particule.

Technical Consideration

Considering the simplicity of the algorithm and what should be done, I tried to write a very simple code.

The code is in c because the algorithm is ultra complex (which is not the opposite of simple) and this should go fast.

I used a table for objects because this go fast, I don't have to add elements and by the way I can do easy copy of the table using memcpy which is very important, it's the fatest c library function...

The input and the output are in Json, so feel free to create any initial condition and to write and to she a renderer (I wont write it, I do not like graphical stuff).

TODO

Short term

• Test and control of the operation
• Changing output file to the format as input (to enbale "to be continued...")

Long term

• Integration of funnier algo like "Particle Mesh"

Miscellaneous

    ╚⊙ ⊙╝
  ╚═(███)═╝
 ╚═(███)═╝
╚═(███)═╝
 ╚═(███)═╝
  ╚═(███)═╝
   ╚═(███)═╝