#stealthmode for a few days as I design and code the core agent #AIsystem. This is one of the big core #gameplaysystems and why this game will be fun to play. The goal is to create a #simulation of government reaction to random events and storyline events which carry the gameplay. The countries will have there own policy objectives and these might align with you or not. You have to make peace or fight back. Allies can turn on you if the conditions change and enemies can become friends. It is composed of two parts a agent brain which stores the configuration for AI behaviors and a list memory of events which take place before and during the game. the events happen over the course of the game vary greatly from a new political group taking power to a terrorist attack, military sales, all the these seeming random events shape how each country reacts to your government. World events which are related to your government are tracked and feed into a #RANSAC or Random Sample conesus #machinelearning #algorithm which is used to separate outliner events, then a separate algorithm called a #Gaussian #Mixturemodel is used to group all your seemingly unrelated events into subpopulations the more dense the subpopulations become, creates thresholds for the decisions trees to determine what goals to carry out.
Ok so what does that mean? the dataset filters out things that it perceives as not important based on the goals at the time. Weighting of those events allows the country AI will choose between science, territory, economic, military goals. A science goal could be to unlock a research tech or achieve a storyline goal like spaceflight.
The route that the country agent will take to achieve this is completely unscripted, the AI agent will determine the best set of decisions to achieve that goal in the time given, it could be research the item, it could be use a Intel operation to steal it, it could be create a trade deal with a country that already has the tech, the AI behaviors will be graded by human testers that will help train the system to make the AI decisions more human.