Once the challenge to the Competitive Complex System has been modeled as an abstract board game, the LG Engine is unleashed within the system. The LG Engine has several unique and revolutionary capabilities that set its performance far ahead of any other artificial intelligence system, regardless of the industry or application.
In order for most AI programs to find solutions, nearly every possibility in the Competitive Complex System must be analyzed. The traditional software approaches are often called “search problems,” “expert systems,” or “brute force” programs. The problem with this approach is that the possibilities within a system usually far outweigh the computational resources.
The Deep Blue chess program that defeated World Chess Champion Gary Kasparov, calculated 200 million possibilities per second using a vast array of chips. Still, the exponential growth in complexity limited it to looking only 8 to 10 moves ahead. Imagine the difficulty the computer would have if the board were 100 x 100, or if the game were played in three dimensions instead of two. Even supercomputers can’t figure out systems of this complexity in real time - the speed at which modern problems occur and demand to be solved.
The LG Engine, on the other hand, doesn’t search at all. Instead, it “thinks” like a specially trained human expert. Masters of strategy make decisions and solve problems by building an understanding of how the forces within their system relate to and interact with each other. Problems are “decomposed” into smaller, more easily handled “zones.” Usually, this is done by figuring out where there is action-reaction within the system, and separating these parts from forces that will not influence this action-reaction.
While keeping in mind what is happening in the other zones (dynamic decomposition), the expert quickly analyzes the tactical or forcing maneuvers in the action-reaction zones and makes an assessment. Then, they integrate the zones together again. Once they have an understanding of how these forces operate in the Competitive Complex System, the solution for how they should proceed is nearly instantaneous. LG works the same way, except with algorithms instead of brain cells.
In fact, comparison tests of solution times for problems with proven optimal paths, the LG Engine was able to produce unheard of performance enhancements. For example, a problem for which a brute force search needs to examine over 1,000,000 positions to solve optimally, the LG Engine needed just 54 positions. On a problem that would take other AI programs 1.23 x 1020 to find the optimal path, the LG Engine needed just 200 nodes. (A computer operating at 100 billion nodes per second would need about 38 years to find a brute force solution for 1.23 x 1020) And these solutions were achieved before a number of refinements to LG were made. We now have the capability, for instance, to analyze these Competitive Complex System with multiple sides moving concurrently (just like in the real-world).
This immensely important advance makes the engine capable of handling challenges as they occur in real life situations, where one side isn’t waiting patiently for the other to make their move. This level of search reduction—from the astronomical to nearly none—enables the LG Engine to pursue applications previously considered impossible to solve.
« Modeling Complex Systems
Linguistic Geometry works on complex systems. Complex systems have many essential components that are metaphorically similar to chess and other strategic games. The board, the pieces, the goals and rules describe the aspects of a Competitive Complex System. Almost any Competitive Complex System can be described with a mathematical representation and made into an abstract game. With this ability to convert Competitive Complex System into a game, the future applications are unlimited. The most important part of this discussion is the fact that this one-to-one modeling of human intelligence has not been accomplished by any other technology—nor is anyone likely to do so as most have given up trying.
The Field of Play
First, a playing space or board is identified. In chess, an 8x8 two-dimensional board is used. LG allows for multidimensional “boards” of any size. The spaces on the board can be of any size or shape. LG has utilized boards that represent power grids, air space, and transportation systems in the past. The board could be represented as a real or imagined. The possibilities are endless, especially for interactive Competitive Complex System.
Once the field of play has been established, the pieces are added. Pieces could be business executives, wall street analysts, bio-pharmacologists, generals, athletes, or almost anything else that could move around and act in a competitive or hostile way. Like in chess, these pieces can be identified as being more or less valuable, they can move or think in certain ways and at slower or faster speeds, and can have capabilities that enable them to protect or destroy other pieces within the game.
The Goals and Rules
In Competitive Complex System, teams in competition with each other are identified, and these teams or sides are given goals to guide their behavior during the game. The strategic element of the games is a result of how the pieces work together to achieve a common goal. In chess, the ultimate goal is checkmate, and any sacrifice that leads to this goal is encouraged.
In war, there may be similar sacrifices, but the commanders must assess whether losing forces is worth attaining the victory. Using this concept it is easy to conceive of "wall street games" of the bulls versus the bears. In games and in real life, constraints allow or disallow certain actions by the players in the field of play. These constraints will be different in each Competitive Complex System, depending on the level of reality or fantasy.
« Competitive Advantage
The LG Engine makes use of the following features to give it real-life capabilities (other AI software engines have none of these capabilities, which provides STILMAN with a long-term advantage over competitors):
Simultaneous moves (multiple pieces can be moved at the same time by each side)
No-Search solutions (lessens the hardware load and expense compared to traditional AI programs)
Various resolution power ( ranging from tactical battles to global strategic planning)
Simultaneous games with hierarchical data feeds (bottom-up and top-down information flow between games of different resolutions)