Unit 3: Intelligent Agents - Practice Quiz

1 What is an intelligent agent?

2 What does an agent architecture provide?

3 What does the letter P represent in the PEAS framework?

4 Which PEAS component allows an agent to collect information from its surroundings?

5 In a robotic vacuum cleaner, which component is an actuator?

6 What does a rational agent generally try to do?

7 Which environment is fully observable?

8 Which type of environment can change while an agent is deciding what to do?

9 How does a simple reflex agent usually select an action?

10 What does a model-based agent maintain?

11 What guides the actions of a goal-based agent?

12 Which activity is commonly used by a goal-based agent to choose actions?

13 What does a utility function represent?

14 What is the main feature of a learning agent?

15 Which part of a learning agent evaluates how well the agent is performing?

16 What is an autonomous intelligent system?

17 What is a multi-agent system?

18 Why do agents coordinate in a multi-agent system?

19 What is the main idea of human-AI collaboration?

20 Which is a common application of an intelligent agent?

21 A warehouse robot receives camera images, updates an internal map, selects a route, and sends commands to its wheels. Which component of the agent architecture is primarily responsible for selecting the route?

22 Two delivery robots use identical agent programs, but one has more accurate sensors and faster motors. Why may their behavior differ in the same environment?

23 For an autonomous taxi, which item belongs to the performance measure in a PEAS description?

24 A crop-monitoring drone detects dry areas and releases water. Which pairing correctly identifies one sensor and one actuator?

25 A medical diagnosis agent recommends the treatment with the highest expected benefit based on incomplete test results. The patient later reacts badly. Does this outcome prove that the agent was irrational?

26 A chess-playing agent can observe the complete board, but it cannot predict the opponent's next move. How is the chess environment best classified?

27 A cleaning robot must choose actions while people continuously move chairs around it. Which property most directly makes planning difficult?

28 A vacuum agent cleans when its current square is dirty and otherwise moves right, without remembering previous locations. What limitation is most likely?

29 A robot can see only the corridor directly ahead but remembers doors observed earlier. Which feature makes it a model-based agent?

30 A route-planning agent knows several roads that reach a destination. What distinguishes its behavior from that of a simple reflex agent?

31 A puzzle-solving agent has a goal test but no preference among solutions. It finds two valid solutions, one requiring 8 moves and another requiring 15 moves. What can be concluded?

32 A delivery agent compares a fast but risky route with a slower, safer route. Which capability specifically supports this trade-off?

33 A recommendation agent initially makes poor suggestions but improves using user ratings. In a learning-agent architecture, which component uses the ratings to evaluate recent behavior?

34 A learning agent keeps recommending familiar products and receives little information about new products. Which component should encourage trying less-certain alternatives?

35 Several independent warehouse robots share narrow aisles and charging stations. Why is this a multi-agent system rather than a single autonomous system?

36 An exploration rover operates for several hours without human commands and adjusts its path after detecting unstable ground. Which property most strongly demonstrates autonomy?

37 Three rescue drones need to search separate zones without duplicating effort. Which mechanism most directly improves coordination?

38 Two network agents repeatedly reserve the same communication channel, detect a conflict, and retry at the same time. Which change is most likely to resolve the repeated conflict?

39 An AI system flags suspicious medical scans and displays the image regions that influenced each flag, while a radiologist makes the final diagnosis. What is the main collaborative benefit?

40 A smart-grid agent predicts electricity demand and shifts flexible appliance usage to off-peak periods. Which outcome best indicates that it is acting effectively?

41 An agent receives percepts through sensors and selects actions through actuators. Which architectural property most directly distinguishes an intelligent agent from a fixed stimulus-response device?

42 For an autonomous ambulance-dispatch agent, which PEAS specification best captures the performance measure?

43 A warehouse robot is evaluated only by the number of packages delivered per hour. It begins taking routes that increase collision risk and damage fragile goods. What is the primary PEAS design failure?

44 An agent must choose between two actions. Action has a 0.7 probability of producing utility 10 and a 0.3 probability of producing utility 0. Action has a 0.4 probability of producing utility 18 and a 0.6 probability of producing utility 2. Which action is rational under expected-utility maximization?

45 A cleaning agent receives an incomplete percept: its current room appears clean, but adjacent rooms are unobserved and may contain dirt. Which characterization best describes its environment?

46 A robot vacuum follows the rule "if the current tile is clean, move forward." It repeatedly crosses a narrow clean corridor and misses a dirty branch. Which modification most directly addresses the failure?

47 A simple reflex agent is operating in an environment that is fully observable, deterministic, static, and finite. Under which condition can it still be inadequate compared with a model-based agent?

48 A navigation agent has two routes to a destination. One route is shorter but passes through a road that may close before arrival; the other is longer but reliable. Why does a goal-based agent alone have difficulty selecting between them?

49 A planning agent has a goal of reaching a clean-energy target. A proposed action reaches the target sooner but permanently disables a safety-monitoring subsystem. Which issue is most directly exposed by relying only on goal satisfaction?

50 A utility-based medical triage agent assigns utilities to outcomes using survival probability, treatment burden, and fairness constraints. A learning component updates these estimates from outcomes. What is the principal benefit of this combination?

51 A learning agent performs well during training but repeatedly fails after a sensor is recalibrated, even though the environment has not changed. Which subsystem should be examined first?

52 An agent chooses actions using , but its learned probability model becomes overconfident after observing only a few examples. What is the most appropriate concern?

53 A fleet of delivery drones shares airspace and independently selects routes. Each drone has a local objective, limited communication, and no central controller. Which property most strongly makes this a multi-agent system rather than merely a distributed implementation?

54 An autonomous underwater vehicle must operate for weeks with intermittent communication, uncertain localization, and changing currents. Which design choice best supports meaningful autonomy?

55 Two warehouse robots independently select the same narrow aisle because each optimizes its own travel time. Which coordination mechanism most directly prevents this conflict?

56 A team of search agents receives a shared reward for locating targets. Agents begin duplicating searches in the same region because individual contribution is not observable. Which intervention best addresses the coordination failure?

57 An AI assistant recommends rejecting a loan application with high confidence, but a human reviewer sees evidence that the applicant's income data are outdated. Which collaboration design is most appropriate?

58 A human supervisor monitors an AI-controlled manufacturing line. The AI requests approval for nearly every low-risk adjustment, causing alert fatigue and delayed responses to rare hazards. Which redesign is best justified?

59 An algorithmic trading agent achieves high returns in historical simulations but performs poorly after market conditions change. Which evaluation flaw is most likely responsible?

60 A hospital bed-allocation agent improves average occupancy but increases waiting time for patients requiring isolation beds. Which change best reflects a careful application design?