RIPR Resource Allocator

Can RIPR be used for decision making in an EW simulation?

Electronic Warfare (EW) can be a complex arena to play in, and RIPR can help. A RIPR agent can exist on any kind of platform and it can perform any kind of platform task. A task to be performed might be as simple as flying towards a point, or maybe even firing a weapon. A common application of the RIPR framework is to make a regular platform into a RIPR agent by adding a RIPR processor to it. The RIPR processor can control the platform’s maneuvering and weapons in order to move towards the right locations and engage the right targets. A platform’s weapons might include missiles, bombs, or in this case: various EW jamming emitters. A RIPR agent can act in conjunction with other other RIPR agents (often through a commander) or it can act solo. Much of the communication between RIPR agents is in the form of job bidding and job allocation; this is described in more detail below. RIPR agents can retrieve data from any aspect of the platform and can control any aspect of the platform. It is typical to make your RIPR agent control some kind of weapon targeting.

Here is an example of a simple platform structure of a RIPR agent:

../_images/ripr_component_interaction.jpg

In a EW jamming simulation, it is reasonable to have a RIPR agent control the jamming weapons and perhaps leave the maneuvering to a pre-established route.

So what is the EW decision making that needs to be done?

A player in a EW jamming simulation commonly needs to answer the question “which targets or threats should I be jamming?”. The decisions to jam some threats and ignore others are important and can determine mission success or failure. The decisions are made difficult when you have many jamming emitters at your disposal and many targets that likely need to be jammed. The decisions are also dynamic, and might need to be changed as the threat environment changes or as the player’s status changes. This situation has just become a dynamic resource allocation problem, a trading game, where a player should attempt to optimize his effectiveness over a period of time.

How can a RIPR agent make these kinds of decisions?

One feature of RIPR is highly tuned to this kind of problem: The Job Board. In order to use a tool such as the Job Board to solve a typical dynamic resource allocation problem, we need to define the problem in terms of Job Board components. The key components behind the Job Board tool are jobs, bidders, and bids. Jobs are tasks to be done, bidders represent some manifestation (perhaps a platform, or even a weapon) that can perform the job, and bids represent an ability to perform said jobs, a measure of effectiveness or capability. A job is known in WSF script as a WsfRIPRJob. Lets break down the EW decision making into these three components; we’ll use the example AIJAM agent along the way.

Jobs - The tasks to be performed in this case are simply jamming of various targets or threats. Typical ways to get targets to jam is to read tracks from the track manager. The track manager might contain tracks received from sensors, received from the network, or it might even contain pre-briefed tracks (representing intel you might have of the targets). You can simply create a “jamming” job from a track like this:

on_update
  foreach (WsfTrack track in PLATFORM.MasterTrackList())
  {
    if (this is a track you want to jam)  //pseudo code
    {
      if (job does not already exists for track)  //pseudo code
      {
        WsfRIPRJob newJob = WsfRIPRJob.Create( PROCESSOR, "jam type", "job description", 1.0, 1);
        newJob.SetData("target name", track.TargetName());
        AddJob(newJob);  //adds the new job to this platforms job board
      }
    }
  }
  ...
end_on_update

Now a jamming job exists on that RIPR agent’s Job Board for each track that platform has awareness of. Anybody can bid on and perform these jobs: subordinates, or even the agent himself. Typically a commander platform creates and owns jobs on it’s Job Board and the subordinates will be making bids to him and they will be tasked with performing the jobs.

Bids - Commanders need to know how well their subordinates can do jobs so that they can effectively distribute the tasks to the proper agents to perform them. A bid represents an agent’s capability to perform a job. Typically, a subordinate agent will query for his commander’s jobs and selectively make bids on them. He should only bid on job types he understands, and he should only bid on jobs he can perform. Here is an example of a subordinate looking for a jamming task and bidding on it:

query_bid
  if (JOB.Name()=="jam type")
  {
    return CalculateEffectivenessAgainst( JOB.GetData("target name") );  //pseudo code
  }
  return 0.0;
end_query_bid

on_update
  foreach (WsfRIPRJob job in GetRIPRCommanderProcessor().GetJobs())
  {
    double bid = QueryBid(job);  //this calls the agents own "query_bid" script block
    job.BidJob(bid);           //here the bid is made on the job
  }
  ...
end_on_update

NOTE - So up this point we have an agent creating jobs to jam threats, and we have an agent bidding on those jobs. The bidding we have exampled here is a gross oversimplification, because a platform might not just have one bid for a particular threat. A platform usually has multiple jammer emitters, and each emitter has multiple beams that can be utilized, and each beam might be modulated (a few spots for each beam) so as to hit different threats at different times. We will cover the more complex case of multiple bids further down. For now, in this simple example, a platform will provide one bid for each job and will only win (or be tasked with) one job to perform.

Performing a Job - As an agent makes bids on jobs it is communicating to the commander “this is my status, and how well I can perform the various tasks you might give to me”. After this communication, the bidding, takes place the agent can then ask its commander “what do you want me to do now, what is my task?”. If it receives a job to perform, it should act on it. Here is an example of retrieving your task and doing something with it:

script void DoJob(WsfRIPRJob aJob)
  if (aJob.IsValid() && aJob.Name()=="jam type")
  {
    if (not already jamming the target)  //pseudo-code
    {
      WsfWeapon jammer = PLATFORM.Weapon("jammer name");
      WsfTrack  target = GetTrack(aJob.GetData("target name"));
      jammer.StartJamming(frequency, bandwidth, target);
    }
  }
end_script

on_update
  ...
  WsfRIPRJob job = GetRIPRCommanderProcessor().GetJobFor(TIME_NOW, PROCESSOR);
  DoJob( job );
  //you might also need some script to stop jamming any previous targets
end_on_update

NOTE - Before this script block, the platform has already bid on his commander’s various jobs. Now he queries for a job, and might receive a job to perform.

But a jamming platform usually jams many things at once!!!

This is fine, RIPR can accommodate. In the simple examples above, the platform was making one bid on each job. His bid represented his total capability to perform the job as a platform. The example script also showed the platform only receiving one job to perform. How does an agent win multiple jobs and perform them? He can do this by using the multi-channel RIPR feature.

Multi-Channel bidding - If a platform can perform multiple jobs in parallel, then he should bid on jobs & win jobs in parallel. One example of such an agent is an ground-based air-defense tracking radar. The tracking radar is sophisticated enough to keep quality track reports on several potential threats. How many threats he can keep track of is a function of his radar. If his radar can handle three targets at once, then this platform should be providing three separate bids for each “track target” job. Each bid represents a “slot” on the radar and that “slot” on the radar provides it’s bids through its own channel. For example:

script_variables
  array<string> channels = array<string>();
end_script_variables

on_init
  channels.PushBack("track target");  //slot 1 of radar
  channels.PushBack("track target");  //slot 2 of radar
  channels.PushBack("track target");  //slot 3 of radar
  SetNumJobChannels(channels.Size());
end_on_init

on_update
  foreach (WsfRIPRJob job in GetRIPRCommanderProcessor().GetJobs()) //bid on any relevant jobs from commander
  {
    // 'index' is also the platform's channel index
    for (int index=0; index < GetNumJobChannels(); index=index+1)   //bid through channels
    {
      string type = channels.Get(index);
      if (type == job.Name())                                       //a way to see if channel is appropriate for job
      {
        double bid = QueryBid(job);    //this calls the agents own "query_bid" script block
        job.BidJob(bid, index); //here the bid is made on the job, for channel 'index'
      }
    }
  }
  for (int index=0; index < GetNumJobChannels(); index=index+1)
  {
    WsfRIPRJob job = cmdr.GetJobFor(TIME_NOW, PROCESSOR, index);
    DoJob(job, index);
  }
  ...
end_on_update

You might ask yourself “why are the channels looped over twice?”. The answer is because you want to make sure every possible bid has been made before you ask the commander who wins the job. So each job is given a bid from each channel, and THEN you can ask what job each channel might have won (maybe no jobs).