Assignment 1 Decision making concepts,State machines and Fuzzy logic

         

Decision making concepts

The first thing that I will talk about in this section of the assignment, under the topic of AI in games, is decision making concepts i.e. the decision tree or the classification tree. Every AI character makes a decision to do something such as the decision to attack the player character if they are seen (Artificial Intelligence, 2010). Each decision is put in to a tree like graph which is called the decision tree and the decision itself is made up of child and parent nodes. To make sure that the tree works properly the child nodes are under the parent nodes and have to represent all the possible decisions that are in play. These may be yes, no, maybe or an action like pick up, store or drop (Gallear, 2017). So with a decision tree it starts off at the root where the decision is made such as yes to take cover, then another decision is made and another until it get to the leaf which is the end node for that section of decisions (Artificial Intelligence, 2010). So the advantages of decision trees is that all of the AI possibilities are show in the trees like a diagram so the AI can weigh up all of the decisions that are put down to make the best possible decision based on the actions that have already been carried out. On the other hand the decision trees are relatively easy to understand but they get a lot more complex as more decisions are made and the outcome is decided and this makes it harder to understand making this a disadvantage (Hamel, 2017) .     

      

(study.com, 2017)

State machines

The second thing that I will talk about in this section of the assignment, under the topic of AI in games, is state machines. These are the 'states' that the AI can be set with. However these AI players can switch between these different 'states' and this normally happens in most games when a character is performing a lot of different actions dependant on the situation at hand. So the 'states' can be things such as wandering, attacking, evading and finding aid, if the AI has been injured. Also, I think that an example of a game that uses this is Assassins Creed Origins because the enemy AI has 'states' that they switch between such as, having a wandering state when the player is not provoking them but if they are provoked they will switch to the attack state and start to attack the player. Occasionally, however, they will switch to the evade state, but there are two other states which are eating if they feel hungry and a sleep state when they feel tired. An advantage of state machine and why they are used in a majority of games is that it makes the AI characters feel like they are actually reacting to what has happened in the game, making the game feel more realistic and therefore more exciting to play by heightening the player experience as well as that key word immersion. A disadvantage is that these are only states and sometimes they will stay on a state such as flee so you have to chase them which can be irritating.                      

(flylib, 2013)

Fuzzy logic

The third thing that I will talk about in this section of the assignment, under the topic of AI in games, is fuzzy logic. Classic example of this is a health bar or stamina bar and logic is normally either alive or dead, but when it comes to fuzzy logic it deals with the concept of partial-truth. This is the bit in between for example if the AI has lost 50% of his life this is the middle ground, the bit in between but it is also where the AI player draws the line. So, if the NPC is low on health they will make the decision to either fight on and die in the battle or they could flee to reserve health. In the Elder Scroll series, Skyrim in particularly, the AI does this exact thing, they either continue to fight and die, flee and hide or flee and heal themselves. One benefit might be that it makes the battles last longer to improve immersion and give you a challenge. It also makes it more realistic. It can be used for simple item and weapon selection or even for the control of enemy units, as well as by you to rank player and NPC's. So as I said before with fuzzy logic the advantage is that it helps create a challenge and makes it feel realistic because you aren't necessarily going to die from one shot. The disadvantage is that it can make it hard on the player as the enemy won't die from one shot and therefore might flee. This means you may have to chase them which could affect the player experience if the player is the type of person who likes one shot one kills and doesn't like to chase the enemies. It could also drop the immersion in their eyes too.                      

(fullofstars, 2007)