Building a 3D Neural Network for Bot AI?

So the Idea

This is supposed to be a neural network in 3d cubes drag and drop at some point...

1...The cube sends out a Pulse (name,name,name,number) call function to each other cube / clone .

2...Each cube checks if the Pulse caries its name.

3...If it true it edits the number and adds the next 3 names.

4... Finally it sends a Pulse to the next script and function as a Pulse(name,name,name,number).

1-4..... this repeats until the pulse hits an output block or fails to pass all thresholds on the way.

Working On / Questions

A?...At the moment its a structure, and only transmits if you build and link each block by hand,
so how to make it link like a chain while instantiating to avoid the by hand part?

B?... How to do drag and drop cubes?

C?... what is wrong with the set parent functions i can not get the links to stay stable....note there is an edited set if instantiate calls that have a glitch in the nerveController script?

E?... what should i change the name of this question to...?

Notes And thank you Phil Rousse, for the awesome first response it is appreciated

Link to package: 77.81 kbs: 4 scripts, 2 folders, 3 pngs, 4 prefabs...very simple is the goal https://www.dropbox.com/s/jykmhw9prj42lh7/Neural%20Cubes.unitypackage

Nerve Controller

 public class nerveController : MonoBehaviour 
 {
     // Nerve cells use to send/recive Pulses
     public inputCell linkInputCell;
     public nerveCell linkNerveCell;
     public outputCell linkOutputCell;
 
     public Transform NerveController;
     public Transform inputCell;
     public Transform nerveCell;
     public Transform outputCell;
 
     public Vector3 grid=new Vector3(9.0f,9.0f,9.0f);
     Vector3 location=new Vector3(0.0f,0.0f,0.0f);
      
     void Awake(){
 
         location= transform.position;
         Instantiate (inputCell, new Vector3 (location.x+1, location.y+0, location.z+1), Quaternion.identity);
         Instantiate (outputCell, new Vector3 (location.x+grid.x, location.y+grid.y+1, location.z+grid.z), Quaternion.identity);
 
         for(int x=1;x<=grid.x;x++){
             for(int y=1;y<=grid.y;y++){
                 for(int z=1;z<=grid.z;z++){
 
                     Instantiate (nerveCell, new Vector3(location.x+x, location.y+y, location.z+z), Quaternion.identity);
 
 
                     //GameObject instance = Instantiate (nerveCell, new Vector3(location.x, location.y, location.z), Quaternion.identity) as GameObject;
 
                     //instance.transform.parent = NerveController;
                 }
             }
         }
         //GameObject instance = Instantiate (outputCell, new Vector3 (location.x+grid.x, location.y+grid.y+1, location.z+grid.z), Quaternion.identity) as GameObject;    
 
         //instance.transform.parent = NerveController;
         
         //GameObject instance = Instantiate (inputCell, new Vector3 (location.x+1, location.y+0, location.z+1), Quaternion.identity) as GameObject;
 
         //instance.transform.parent = NerveController;
 
     }
 }

Nerve Cells

 public class nerveCell : MonoBehaviour
 {
     // Nerve cells use to send/recive Pulses
     public inputCell linkInputCell;
     public nerveCell linkNerveCell;
     public outputCell linkOutputCell;
         
     public float output = 0;
     public float multiplyer;
     public float threshhold;
 
 
     public Vector3 Name = new Vector3 (0f, 0f, 0f);
     Vector3 callNamea = new Vector3 (1f, 0f, 0f);
     Vector3 callNameb = new Vector3 (0f, 1f, 0f);
     Vector3 callNamec = new Vector3 (0f, 0f, 1f);
 
     // awake so choose its settings for multi and thresh
     float awake = 0; 
     void Awake ()
     {
         if (awake == 0) {     
             //update own name
             Name = transform.position;
             multiplyer = Random.Range (-1.5f, 2.5f);
             threshhold = Random.Range (-1.5f, 2.5f);
             awake++;
         }
     }
 
     public void Pulse (float pulseIntensity, Vector3 callNamex, Vector3 callNamey, Vector3 callNamez)
     {
         //
         if (pulseIntensity * multiplyer >= threshhold) {
 
             //update own name
             Name = transform.position;
 
             //check if Pulse is ment for self or else ignore
             if (Name == callNamex || Name == callNamey || Name == callNamez) {
 
             //Update the Pulseintensity before sending
             pulseIntensity = pulseIntensity * multiplyer;
             output = pulseIntensity;
 
             //update the call names before sending
             callNamex = callNamea + Name;
             callNamey = callNameb + Name;
             callNamez = callNamec + Name;
 
             // now send updated pulse to next nerves called by location
             linkNerveCell.Pulse (pulseIntensity, callNamex, callNamey, callNamez);
             linkOutputCell.Pulse (pulseIntensity, callNamex, callNamey, callNamez);
 
             }
         }
     }

}

Input Cell

 public class inputCell : MonoBehaviour
 {
     // Nerve cells use to send/recive Pulses
     public inputCell linkInputCell;
     public nerveCell linkNerveCell;
     public outputCell linkOutputCell;
 
     //this is visibal while in play so to adjust the settint
     public float inputIntensity = 3;
     public float output = 0; 
 
     //These store the cells own name and the names of those it talks to....vector phone numbers
     public Vector3 Name = new Vector3 (0f, 0f, 0f);
     Vector3 callNamea = new Vector3 (1f, 0f, 0f);
     Vector3 callNameb = new Vector3 (0f, 1f, 0f);
     Vector3 callNamec = new Vector3 (0f, 0f, 1f);
     
     //The Function to fire data in to the network
     void Pulse (float pulseIntensity)
     {
             //let us check if pulse came in...
             output = pulseIntensity;
 
             //update own name
             Name = transform.position;
 
             //make the call names before sending
             Vector3 callNamex = callNamea + Name;
             Vector3    callNamey = callNameb + Name;
             Vector3 callNamez = callNamec + Name;
 
             // now send updated pulse to next nerves called by location
             linkNerveCell.Pulse(pulseIntensity, callNamex, callNamey, callNamez);
             linkOutputCell.Pulse(pulseIntensity, callNamex, callNamey, callNamez);            
     }
     //        NullReferenceException: Object reference not set to an instance of an object inputCell.Pulse (Single pulseIntensity) (at Assets/xyz Nerve System/inputCell.cs:37) inputCell.OnGUI () (at Assets/xyz Nerve System/inputCell.cs:48)
 
 //A TEST button
 
         void OnGUI ()
         {
                 // make a button that fires a pulse out of the input node in to the network
                 if (GUI.Button (new Rect (10, 10, 60, 30), "Pulse")) {
                         Pulse(inputIntensity);
 
 //    SendMessage("Pulse",inputIntensity,callNamex,callNamey,callNamez);
 
 //Pulse(inputIntensity,callNamex,callNamey,callNamez);
 
 //Pulse(0+1);
                 }
 
         }
 
 }

Output Cell

 public class outputCell : MonoBehaviour 
 {
     // Nerve cells use to send/recive Pulses
     public inputCell linkInputCell;
     public nerveCell linkNerveCell;
     public outputCell linkOutputCell;
 
     //make shure  the Cell only wakes up once
     private float awake = 0; 
     public float output=0;
     
     // when created and wake up takes a  name from its position, decides on output targets
     public Vector3 Name = new Vector3(0f,0f,0f);
     Vector3 callNamea = new Vector3(1f,0f,0f);
     Vector3 callNameb = new Vector3(0f,1f,0f);
     Vector3 callNamec = new Vector3(0f,0f,1f);
 
     // then and choose its settings for multi and thresh
     void Awake(){
         if (awake==0){
             Name=transform.position;
             awake++;
         }
     }
     
     public void Pulse(float pulseIntensity, Vector3 callNamex,Vector3 callNamey, Vector3 callNamez){
 
         //update own name
         Name = transform.position;
 
         if(Name==callNamex || Name==callNamey || Name==callNamez){
             output=pulseIntensity;
         }
     }
 
     void OnGUI() {
         GUI.Button(new Rect(80, 10, 50, 50), "#"+output);    
     }
     
     
 }
 

suggestions on the code would be awesome.

This is supposed to resemble a neural network in 3d

thanks for any and all help

feel free to learn, i wrote it to share....lol : )/