WEEK 1

Question 1

What is decision making ?

Answer 1

it is the ability of a character to decide what to do

Question 2

what are the main parts of making a decision

Answer 2

1. input
2. generate of
3. output

Question 3

what is the output of a decision ?

Answer 3

External Changes

internal changes

Question 4

what is the input of decision maker ?

Answer 4

external knowledge

internal knowledge

Question 5

What is a decision Tree

Answer 5

are simple structures, quite fast and easy to implement and
understand. The concept is simple: given a set of input information, decide
what’s the best action to execute.

Question 6

What are decision tree’s made up from ?

Answer 6

nodes and leafs, root

Question 7

what is a decision tree node?

Answer 7

nodes are where the decision are made, each node has a limit of 2

Question 8

what is the root ?

Answer 8

the root is the first node of the decision tree,

Question 9

what are the leafs ?

Answer 9

at the leafs a action is attached so that the agent does the action that the leafs says

Question 10

how does the nodes makes a decision ?

Answer 10

Decisions at each node are made based on the internal knowledge of the agent,
or queried to a centralised manager object that provides this information, carried out automatically, leaves can have the same

Question 11

what type of decision do the nodes make ?

Answer 11

typically they check a single numeric value an enumeration, or a single
boolean condition.
but you can have more complex operations, such as calculating distances, visibility checks, pathfinding, etc

Question 12

How to do and and or in clauses ?

Answer 12

AND
1 1 = 1
1 0 = 0
0 =0

OR
1 = 1
0 1 = 1
0 0 = 0

Question 13

what is a binary decision tree?

Answer 13

The N-ary tree requires less
comparisons: it’s more efficient.But, binary trees are more
common. Efficiency gain is lost
at the code level (sequence of if
statements). Also, binary
decision trees are more easily
optimized.

Question 14

what is a balanced tree

Answer 14

where the tree has the same number of leaves on each branch, o(log2(n))

Question 15

what is the auto time if iit is a total unbalanced tree ?

Answer 15

o(n)

Question 16

other info

Answer 16

However, the optimal tree
structure depends on the
decision nodes. Most
commonly used checks should
be placed close to the root
node, and the most expensive
ones should be located closer
to the leaves.

Question 17

what is aDAG

Answer 17

Directed Acyclic Graph

Question 18

what is random decision tree

Answer 18

it is a descion tree that allows variatiion in the behaviiour by introducing a random element

Question 19

what is HFSM ?

Answer 19

Hierarchical Finite State Machine

Question 20

can you combine FMS’S and Decision Trees ?

Answer 20

yes

Question 21

what are the advantages of BT’s

Answer 21

• Can incorporate numerous concerns such as path finding and planning
• Modular and scalable
• Easy to develop, even for non-technical developers
• Can use GUIs for easy creation and manipulation of BTs

Question 22

a basic BT consists of how many and what nodes

Answer 22

• Conditions: test some property of the game
• Actions: alter the state of the game; they usually succeed
• Composites: collections of child tasks (conditions, actions, composites)

Question 23

talk about selector and sequence

Answer 23

(?)selector is a OR statement

(->)sequence is and AND

Question 24

what is a blackboard ?

Answer 24

witch a blackboard you are decouple the data from tree (rather that, say, use
arguments to the actions).

Question 25

Reuse of Behaviours ?

Answer 25

It is apparent that BTs, if designed properly, can be quite modular and hence it makes sense to re-use whole or partial trees. Actions and conditions can be parameterised and data can be exchanged via the blackboard. This allows one to re-use trees if agents require identical behaviours. This can be done in numerous different ways: • Use a cloning operation • Create a behaviour tree library • Use a dictionary and reference names to retrieve trees/sub-trees (this is normally known as look-up table)

Question 26

what is a goal ?

Answer 26

The character can have one or multiple goals (motives) to accomplish. Not all of them need to be active at the same time, and they most likely will have different priorities.

Question 27

what is a insistence

Answer 27

The insistence of a goal determines how important is to fulfil that goal. The agent will try to accomplish goals with higher insistence or reduce its value

Question 28

what is an actiions

Answer 28

Actions: moves or interactions with the world that serve the agent to accomplish the goals (in several states from now).

Question 29

how do you measure this

Answer 29

Each action can have a measure or utility of how much it affects the insistence of a goal.

Question 30

ex

Answer 30

Example: our character is hungry. • The goal would be to minimize hunger. • The insistence would increase as time goes by without eating. • Actions could be direct (eat) or indirect (pre-heat oven).

Question 31

what are the produmes with GOB

Answer 31

This is a simple and fast approach that generally gives good results. However: • It fails to account for side-effects an action might have. • It does not incorporate timing information

Question 32

What is discontent ?

Answer 32

instead of focusing on reducing the goal(s) insistence, tries to reduce an overall discontent. Discontent can be calculated differently:

Question 33

what are the different ways of calculated the differently

Answer 33

* Simplest approach: sum the insistence of all goals. | * Better: sum the squared value of the insistence of all goals.

Question 34

what is an ex

Answer 34

* Goal 1: Eat (Insistence: 4) * Goal 2: Bathroom (Insistence: 3) * Action: Drink Soda (Utility: Eat - 2; Bathroom + 3) * Action: Visit Bathroom (Utility: Bathroom - 4) In our example: • Action: Drink Soda (Utility: Eat - 2; Bathroom + 3) • If taken: Eat = 2, Bathroom = 6, so discontent (sum: 8, squared: 40). • Action: Visit Bathroom (Utility: Bathroom - 4) • If taken: Eat = 4, Bathroom = 0, so discontent (sum: 4, squared: 16).

Question 35

what is GOAP ?

Answer 35

Goal oriented action planning

Question 36

what does GOAP do?

Answer 36

In order to be able to use GOAP, we need to have a predictive model (a.k.a. forward model) of the world. The agent has an internal simulator that allows to sample what would happen if certain actions are taken from an original state. Then, at each step, we would know which actions are available in our action space

Question 37

what is Knowledge Representation

Answer 37

Knowledge representation: Representing information about the world in a form that an agent can utilize to solve complex tasks.

Question 38

what are the different types of data represent

Answer 38

Symbolic: Specific values (speed=15.0, temperature=30) | • Fuzzy: Fuzzy values (speed=’fast’, temperature=’high’)

Question 39

what does this data refer to

Answer 39

the world, pathfinding (naviigational mesh ), poittioning systems, liine of sight iinformatiion --internal, facts about the world

Question 40

what are the ways of representing this information ?

Answer 40

Simple: arrays, dictionaries / hash maps. • Working memory. • Blackboards.

Question 41

what are the different between long-term memory and short-term memory

Answer 41

• Long-term memory: long term facts (I was attacked by this player, this agent is my friend). • Short-term memory: volatile facts (there’s an enemy behind that door, there’s no food in the fridge).

Question 42

what is the restion for have a working memorie

Answer 42

Working memories function as an internal model of the world, that allows the agent to draw beliefs which might be or not accurate with respect to the real world.

Question 43

what are the advantages

Answer 43

Provide a simple overview to the agent’s state. • Possible place to store agent’s global state variables. • The system is modular.

Question 44

what are the disadvantages ?

Answer 44

Scales poorly for multiple agents. • Centralized data. • Hard to keep track of its usage: what’s modified by whom and when?

Question 45

what is a blackboard made up from

Answer 45

1. Blackboard 2. Knowleadge Sources (KSs) 3. Arbitrer

Question 46

what iis a Knowledge Sources (KSs)

Answer 46

Set of components that are able to operate on the information held by the blackboard. Each one of them operates in certain circumstances (triggered by preconditions) and are responsible for manipulating a specific subset of data in the blackboard. These manipulations could take the form of inserting new facts, changing beliefs, setting new goals and communicating to other KSs only via the blackboard.

Question 47

what are arbitrer ?

Answer 47

: In case of conflicts, the Arbitrer decides which KSs should operate. Different strategies are possible, and they may take into account the relevance of each KS, and their contribution towards goals or sub-goals.