The network is made of nodes and links connecting the nodes which form a network, and these all inherit from the dynamic base interface. The network of nodes and links we shall from this point on call an action network. There evaluator which is an object that transitions to the connected nodes by using the attributes of each node as input and the attributes of the evaluator as input as well, and there is a single output. This output is related to how much the node is regarded from the rest of the possible nodes when choosing the correct node.
You can place a visitor for eating in the evaluator where to be passed to a link, and a visitor for eating while the attributes for the location and amount of food is stored for example. Then all behavior can be resolved if you perform one action at a time. Also, more than one evaluator can transition a network at any given time. This way an entire cognitive process can be constructed within a network.
Networks of nodes will be constructed to produce patterns of behavior that I hope we can eventually make generative. I want to make the network a cognitive network, so we might need to have to address the structure of the network as well. Also, the training set will need to be addressed for certain. So, there is much work to do. At the end of the project I wish to have a development package that can be used for generative cognitive behavioral networks to produce human like behavior.
With your permission, GameDev.net uses analytics cookies to understand how people use the platform.
You can accept analytics or continue with necessary cookies only.
Learn more
Discussion