A
Short History of Distributed Search
Jeff
Cares, Alidade Consulting
I
will talk about what an Information Age Combat Model might look like.
I will discuss traditional models and where we are today and where we
are going.
Click
image for enlargement
If
we start with a look at Iwo Jima during WW II, we can see ships,
people, and lots of boxes. We see the massive scale of equipment
that has to be transported by the military.
Next,
if we look at a photo from the First Gulf War. Again you can see
a photo of a person to show you the scale of equipment needed. You
can see there are more electronics than 1945. We still show up with
a lot of stuff for combat.
Click image to see video
Now
we will contrast both of the these views with a image of a network map
from Kosovo in 1999. This is a snap shot of Internet activities that occurred
everyday at the same time during the bombing. It represents a system that
is evolving. We do not know any more information about this image. We
do not know what the red lines or blue lines represent.
The
question then becomes about what decisions should we make from this information.
This model demonstrates how difficult it is to watch a distributed force
and be able to make decisions. We need to have logic behind how systems
work and discern the advantages to make decisions and policies. If we
are an adversaries just watching, it it is hard to know what we should
do with information. There are behaviors, patterns and human acts happening
that we need to know how to discern.
Simplistic
Physical Model
In
the Simplistic Physical Model, we have three layers. The bottom layer
is all the these targets out there and we blanket on the nest layer a
sensor grid and then put a weapons grid on top of that. We then have to
understand the tactics. We do not want just big "blankets of capability"
out there as our only approach. We need to know about the individual paths
between all of these things. What is the arrangement of sensors and the
arrangement of weapons? We need to understand that when a sensor is trying
to engage it needs to have a better arrangement of the weapons in relationship
to the target. We need to understand the paths through this network or
system.
Click image for enlargement
In
this work, we want to take the abstract and apply mathematics to it. The
goal is to be able to make decisions as well as collect data. There has
been an evolution of mathematics applied to the Navy world. An example
of this is when you look at Combat Theory. There are a few equations out
there that we have used to give us some insights.
(Insert Combat Theory Development & Insights Slide)
The
Hughes Salvo Equation works if the forces are homogenous and are applied
to each other in a stable way. We then played out this scenario to see
if it matched the equation. We were losing up to 30% of combat power in
the exchange. Each missile we shoot has 4 outcomes hits, misses, shot
down or not shot down. We would shoot 12 missiles at one target and we
would loose 30% of our combat power.
How
do we make sure that we shoot perfect shots? Are perfect shots globally
perfect ? Are they perfect for the system, defender or shooter? Very few
interactions are perfect. It is hard to think about 30% waste in a perfect
system. And now think about the challenges we have in distributed forces.
What
if things were heterogeneous? If things were heterogeneous, you would
need a sophisticated map of the parameters. This is difficult because
the parameters change all the time. A rating system will be necessary.
We have to rate certain ships. If we are loosing 30% in the simple exchange,
what do we loose in the heterogeneous equation?
A
great paper by Keith Ho looked at the parameters and developed alternative
calculations. One alternative was to shoot down half a missile. In this
world of missile exchanges, you can not shoot half a missile. Closed form
equations help explorethe dynamics but they can not capture all the complexity
of the interactions. The parameters are dynamic as well as disputed.
What
is a Military Force?
In a military force we have a passive target, autonomous sensor, simple
decider, and autonomous influence.
Click
image for enlargement
In
competing networks, there are many ways they can interact.
Click
image for enlargement
The
Warfare Model has 36 dimensions of all the ways they can usefully interact.
An example of an interaction is when my own weapon shoots at my own target
by accident.
Click
image for enlargement
We
can then map this a graph as an Adjacency Matrix. An arrow from the column
element to the row element carries the value of one.
Click
image for enlargement
The
Adjacency Matrix can be used to define cycles. There are eight control
cycles. These are cycles where we control our own sensor and target arrangements.
Click
image for enlargement
There are 50 Catalytic Control Cycles. One equation is where
we control our own sensor arrangements based on our own target arrangements.
The other equation is were we control our own influence arrangements based
on our own sensor and target arrangements.
Click
image for enlargement
There
are 4,950 Competitive Catalytic Cycles.
Click
image for enlargement
There are 12 Combat Cycles where we can influence the opposition
targets.
Click
image for enlargement
Questions
Why are there more control cycles than combat cycles? What
are the key centers of gravity? What are the statistical mechanics of
combat?
What you are doing is identifying all the possible paths.
A formulation could be done to yield the results of the others. What are
the mechanics that make it happen? What are all the paths that could occur
in the process? Who found what when so we know when they found it so we
can know when to shoot it? Most of the effort is in the the arrangement,
but you only get credit for shooting. The intelligence person, who helped
to locate the target, does not get the credit. How do command and control
arrangements contribute to combat?
In the combat cycles is it possible to tell what the independence
is of all those paths? What paths actually produce results? How can we
create more autonomous systems?
Should we add another circle that represents someone who
is supposed to help the decision maker or sensor person? Where do we sever
some cycles so they do not effect the other parts of the system?
|
