Task decomposition strategies in Agentic AI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Task Decomposition in Agentic AI

Which of the following best describes the purpose of task decomposition in agentic AI systems?

A Randomly assigning tasks to agents without any hierarchy or structure

B Combining multiple simple tasks into one large task to reduce computation time

C Eliminating subtasks to simplify the overall task and reduce errors

D Breaking a complex task into smaller, manageable subtasks to improve problem-solving efficiency

💡 Hint

Think about how big problems are easier to solve when split into smaller parts.

Task decomposition helps by breaking down complex problems into smaller subtasks, making it easier for AI agents to solve each part efficiently.

❓ Model Choice

intermediate

Choosing a Decomposition Strategy for Sequential Tasks

You have a task that requires completing steps in a strict order. Which decomposition strategy suits this best?

A Hierarchical decomposition, where subtasks are arranged in a sequence with dependencies

B Parallel decomposition, where subtasks run simultaneously without order

C Random decomposition, where subtasks are assigned randomly to agents

D No decomposition, solving the entire task as one block

💡 Hint

Consider the importance of order and dependencies between subtasks.

Hierarchical decomposition organizes subtasks in a sequence respecting dependencies, ideal for ordered tasks.

❓ Predict Output

advanced

Output of Recursive Task Decomposition Function

What is the output of the following Python function that decomposes a task recursively?

Agentic AI

def decompose(task, depth=0):
    if depth == 2:
        return [task]
    else:
        return [task + '_sub1'] + decompose(task + '_sub2', depth + 1)

result = decompose('task')
print(result)

A ['task_sub1', 'task_sub2_sub1', 'task_sub2_sub2']

B ['task_sub1', 'task_sub2_sub1']

C ['task_sub1', 'task_sub2_sub1', 'task_sub2_sub2', 'task_sub2_sub2_sub1']

D ['task_sub1']

💡 Hint

Trace the recursive calls and note when the base case stops recursion.

The function adds '_sub1' at each level and recurses with '_sub2' until depth 2, producing three items.

❓ Metrics

advanced

Evaluating Efficiency of Task Decomposition

You measure the time taken to complete a task with and without decomposition. Which metric best shows improvement due to decomposition?

A Number of subtasks created regardless of time

B Memory usage increase after decomposition

C Decrease in average time per subtask compared to whole task time

D Increase in total computation time after decomposition

💡 Hint

Think about how breaking tasks affects time efficiency.

Decomposition is efficient if subtasks take less time on average than the whole task, showing better manageability.

🔧 Debug

expert

Debugging Task Decomposition Logic in Agentic AI

Given the following pseudocode for decomposing tasks, which line causes a logical error that prevents proper subtask creation?

1: function decompose(task):
2:   subtasks = []
3:   if task is simple:
4:     return [task]
5:   for sub in task.subtasks:
6:     subtasks.append(decompose(task))
7:   return subtasks

A Line 2: Initializing subtasks as an empty list

B Line 6: Recursively decomposing the same task instead of the subtask

C Line 3: Incorrect condition to check if task is simple

D Line 7: Returning subtasks instead of the original task

💡 Hint

Look carefully at what is passed to the recursive call inside the loop.

Line 6 calls decompose(task) repeatedly instead of decompose(sub), causing infinite recursion or wrong results.