Editing Inverse Monte Carlo
Jump to navigation
Jump to search
The edit can be undone. Please check the comparison below to verify that this is what you want to do, and then publish the changes below to finish undoing the edit.
Latest revision | Your text | ||
Line 1: | Line 1: | ||
'''Inverse Monte Carlo''' refers to the numerical techniques to solve the so-called inverse problem in fluids. | '''Inverse Monte Carlo''' refers to the numerical techniques to solve the so-called inverse problem in fluids. | ||
Given the structural information (distribution functions) the inverse | Given the structural information (distribution functions) the inverse Monte | ||
More information can be found in the review by Gergely Tóth ( | Carlo technique tries to compute the corresponding interaction potential. | ||
More information can be found in the review by Gergely Tóth (See the references) | |||
==Uniqueness theorem== | |||
The uniqueness theorem is due to Henderson (Ref. 3). | |||
== An inverse Monte Carlo algorithm using a [[Wang-Landau method|Wang-Landau]]-like algorithm == | == An inverse Monte Carlo algorithm using a [[Wang-Landau method|Wang-Landau]]-like algorithm == | ||
==== Procedure === | A detailed explanation of the procedure can be found in reference 1. A sketchy description for a simple | ||
The simulation procedure is divided | fluid system is given below: | ||
the radial | |||
=== Input information === | |||
*''Experimental'' Radial distribution function <math> g_0(r) </math> at given conditions of temperature, <math> T </math> and density <math> \rho </math> | |||
*Initial guess for the effective interaction (pair) potential; | |||
: <math> \beta \Phi (r) \equiv \frac{ \Phi(r) }{ k_B T} </math> | |||
=== Procedure === | |||
* The simulation procedure is divided in several stages | |||
* The effective interaction is modified through the simulation in each stage, <math> s </math>, to bias the current result of | |||
the radial distrbution function, <math> g_{inst}(r) </math> to the target <math> g_0(r) </math> by using: | |||
: <math> \beta \Phi^{new}(r) = \beta \Phi^{old}(r) + \left[ g_{inst}(r) - g_0(r) \right] \lambda_s </math>, | |||
where <math> \lambda_s </math> is greater than zero and depends on the stage <math> s </math> | |||
* The simulation in each stage proceeds until some convergence criterium (that takes into account | |||
the precission of the values of <math> g_0(r) </math>) for the global result of the | |||
radial distribution function over the stage, is achieved (See Ref. 1) | |||
* When the simulation on one stage is finished a new stage starts with a smaller value of <math> \lambda </math>: | |||
: <math> | : <math> \lambda_{s+1} = \alpha \lambda_s </math> with: <math> 0 < \alpha < 1 </math> | ||
At the final | * At the final stages, with <math> \lambda </math> being small enough, we can get an effective pair potential compatible with the input <math> g_0(r) </math> | ||
== References == | == References == |