xieguigang 10 years ago
commit cf292cb4fd

Binary file not shown.

After

Width:  |  Height:  |  Size: 56 KiB

@ -1,2 +1,12 @@
## Distribute Computing Demo: Genetics Algorithm
### The speed limitation step in GA estimates
One of the speed limitation step in GA eastimates is the fitness sorts for selects the best population that can evolved. For calculate the fitness of a parameter set, ODEs will be computed and the Euclidean distance will be calculated between the estimates function and observation data. This calculation operation will takes a long time to run if the data resolution is very high and population is very large.
For a GA estimation that running on a single machine, the only way to boost this operation is using parallel linq, but a single machine is too power limited, so that a server cluster will required for solving this large scale data analysis problem. For solving a system parameter estimates problem using GA, in a server cluster, then we needs distribute computing.
For running GA on a single machine, makes it run on a server cluster, we just needs devide its population into sevral parts, and then push the fitness calculation task onto each single machine in our server cluster, then we can makes this GA analysis running in distribute method.
![](./Mutation-and-fitness.png)
> A basic calculation in GA

@ -1,15 +1,20 @@
# Microsoft.VisualBasic.Parallel
> Further reading: [Parallel library of GCModeller parallel computing](http://www.codeproject.com/Articles/1076209/Easy-Distribution-Computing-in-VisualBasic)
# sciBASIC# ComputingServices Parallel library
> Further reading on codeproject article: [Easy Distribution Computing in VisualBasic](http://www.codeproject.com/Articles/1076209/Easy-Distribution-Computing-in-VisualBasic)
##### Runtime
Require of VisualBasic server CLI runtime
+ Require of sciBASIC# runtime
> PM> Install-Package VB_AppFramework
```bash
# https://github.com/xieguigang/sciBASIC#
PM> Install-Package sciBASIC -Pre
```
Or reference to source code project:
+ And also required of SMRUCC.WebCloud web runtime for RQL module
> https://github.com/xieguigang/VisualBasic_AppFramework
```bash
# https://github.com/SMRUCC/SMRUCC.WebCloud
```
##### Steps of the distributed computing in this library
@ -24,9 +29,10 @@ Through these steps, your function on the local client should be running on the
**Important NOTE:**
+ Please notices that the function which is running on the remote machine should be a statics method and donot reference to the module variable as the module variable is probably not initialized on the remote machine, just using the variable limits in your function inner local variable.
Example:
```vb.net
```vbnet
Module Test1
Dim m_var As Integer
@ -53,7 +59,7 @@ End Module
### Usage of the remote linq script to query remote resource
```vb.net
```vbnet
Imports <namespace>
var source = http://linq.gcmodeller.org/kegg/ssdb/nucl/xcb
@ -69,9 +75,10 @@ Unlike the SQL langauge, the RQL language is a object-oriented language for quer
#####Using RQL language rest API
First the repository server should implements the RQL services on the server program, the RQL services which is available in the RQL project;
Then your client user that can query your repository server directly by using linq script in two ways, examples as:
###### 1. query repository throw linq script
```vb.net
```vbnet
Imports GCModeller.RQL
var source = "http://linq.gcmodeller.org/kegg/ssdb"

Loading…
Cancel
Save