Section 4 : Online search agents and unknown environments

Commentary

Section Goals

• To identify the online search algorithms needed to solve exploration problems.
• To compare the work principle of online search agents to offline search agents.

Learning Objectives

Learning Objective 1

• Differentiate between online search and offline search.
• Define the following concepts or terms:
  • Exploration problem
  • Online local search
  • LRTA*
  • Random walk
• Analyse the principle of online search agent, and compare it to the classic search process.

Objective Readings

Required readings:

Online Search Agents (see Section 4.5 of AIMA3ed)

Supplemental Readings

Bulitko, V., Lustrek, M., Schaeffer, J., Bjornsson, Y., and Sigmundarson, S. (2008). Dynamic Control in Real-Time Heuristic Search. Journal of Artificial Intelligence Research, 32, 419-452.

Objective Questions

• What are the differences between offline search and online search? Use examples to illustrate your answer.
• Is there a general exploration strategy to solve an online search version based on the relevant offline search?

Objective Activities

• Analyse the supplemental readings for recent research relating to real-time A*. This work can be done jointly with Section 2 of this unit.
• Explore some recent papers on online search from journals such as Artificial Intelligence, Journal of Artificial Intelligence Research, and Computational Intelligence. Discuss the problems, recent progress, and the future direction through the online conference.
• Explore the following online search-related programs from the textbook's website.
  • Online-DFS-Agent
  • LRTA*-Agent
• Complete Exercise 4.14 of AIMA3ed.