For designing a maze problem solver intelligent agent that could act rationally, the rationality at any given time depends on the PEAS. Define the PEAS for this maze problem solver:

Performance measure?

Environment? Complete the following form.

Actuators: Sensors$$Environment$$Observable$$Deterministic$$Episodic$$Static$$Discrete$$Agents$$Q$3$:

There are several agent architectures, which are as follows:

$$ Table$-$driven agents

$$ reflex

$$ reflex with state

$$ goal$-$based

$$ utility$-$based

$$ learning

Which agent architecture would you like to use? Give your reasons.

Q$4$:

Convert the given maze into a graph. $($Hint: For simplicity$$s sake, use those specified grids $($with the alphabet$)$ to name the nodes; their distance is the number of grids between the paired named nodes as their distance. For example, nodes C and R are reachable by forwarding $6$ grids from C towards R or R towards C$.)$

Q$5$:

Considering a portion of the graph for the yellow$-$colored part of the maze, design a state search space in terms of states, state$-$actions, forward actions, and directed edges.Develop a necessary and sufficient theory to claim annotations, A$,$ B$,...,$ Z$,$ a$,$ b$,...,$ z$.$ Obtain a connected undirected weighted graph from a given maze. Will this obtained graph uniquely correspond to the given maze?

Design and implement the agent $($the maze problem$-$solver, an agent, a program, or a simulator$)$ such that for any maze, the agent can find an exit efficiently once it enters the maze. Input to the agent is any maze. $($the agent can take different mazes as its input$).$ The agent consisting algorithms that traverse a given maze to find an exit.