Abstract
Since best‐first search algorithms such as A* require large amounts of memory, they sometimes cannot run to completion, even on problem instances of moderate size. This problem has led to the development of limited‐memory search algorithms, of which the best known is IDA*. This paper presents the following results about IDA* and related algorithms:
1) The analysis of asymptotic optimality for IDA* in [R.E. Korf, Optimal path finding algorithms, in: Search in Artificial Intelligence, eds. L. Kanal and V. Kumar (Springer‐Verlag, 1988) pp. 200-222] is incorrect. There are trees satisfying the asymptotic optimality conditions given in [R.E. Korf, Optimal path finding algorithms, in: Search in Artificial Intelligence, eds. L. Kanal and V. Kumar (Springer‐Verlag, 1988) pp. 200-222] for which IDA* is not asymptotically optimal.
2) To correct the above problem, we state and prove necessary and sufficient conditions for asymptotic optimality of IDA* on trees. On trees not satisfying our conditions, we show that no best‐first limited‐memory search algorithm can be asymptotically optimal.
3) On graphs, IDA* can perform quite poorly. In particular, there are graphs on which IDA* does Ω(22N) node expansions where N is the number of nodes expanded by A*.
Similar content being viewed by others
References
A. Bagchi and A. Mahanti, Search algorithms under different kinds of heuristics — a comparative study, JACM 30(1) (1983) 1–21.
A. Bagchi and A. Mahanti, Three approaches to heuristic search in networks, JACM 32(1) (1985) 1–27.
P.P. Chakrabarti, S. Ghosh, A. Acharya and S.C. De Sarkar, Heuristic search in restricted memory, Artificial Intelligence 47 (1989) 197–221.
R. Dechter and J. Pearl, Generalized best-first search strategies and the optimality of A*, JACM 32(3) (1985) 505–536.
M. Evett, J. Hendler, A. Mahanti and D. Nau, PRA*: A memory-limited heuristic search procedure for the connection machine, in: Frontiers '90: Frontiers of Massively Parallel Computation (1990).
S. Ghosh, A. Mahanti and D.S. Nau, ITS: An efficient limited-memory heuristic tree search algorithm, in: AAAI 1994 (1994) pp. 1353–1358.
P.E. Hart, N.J. Nilsson and B. Raphael, A formal basis for the heuristic determination of minimum cost paths, IEEE Transactions on Systems Sciences and Cybernetics (1968) pp. 1556–1562.
R. Korf, Linear-space best-first search, Artificial Intelligence 62 (1993) 41–78.
R.E. Korf, Depth first iterative deepening: An optimal admissible tree search, Artificial Intelligence 27 (1985) 97–109.
R.E. Korf, Optimal path finding algorithms, in: Search in Artificial Intelligence, eds. L. Kanal and V. Kumar (Springer-Verlag, 1988) pp. 200–222.
J.D. Little, K.G. Murty, D.W. Sweeney and C. Karel, An algorithm for the traveling salesman problem, Operations Research 11 (1963) 972–989.
A. Mahanti, S. Ghosh, D.S. Nau, A.K. Pal and L.N. Kanal, Performance of IDA* on trees and graphs, in: AAAI 1992 (1992) pp. 539–544.
A. Mahanti and A.K. Pal, Worst-case time complexity of IDA*, in: Tenth International Conference in Computer Science (Santiago, Chile, July 1990).
A. Mahanti and K. Ray, Network search algorithms with modifiable heuristics, in: Search in Artificial Intelligence, eds. L. Kanal and V. Kumar (Springer-Verlag, 1988) pp. 200–222.
A. Martelli, On the complexity of admissible search algorithms, Artificial Intelligence 8 (1977) 1–13.
L. Mero, A heuristic search algorithm with modifiable estimate, Artificial Intelligence 23 (1984) 13–27.
N.J. Nilsson, Principles of Artificial Intelligence (Tioga, Palo Alto, 1980).
B.G. Patrick, M. Almulla and M.M. Newborn, An upper bound on the complexity of iterative-deepening-A*, Annals of Mathematics and Artificial Intelligence 5 (1992) 265–278.
J. Pearl, Heuristics (Addison-Wesley, Reading, MA, 1984).
S. Russell, Efficient memory-bounded search methods, in: ECAI-1992 (Vienna, Austria, 1992).
A. Sen and A. Bagchi, Fast recursive formulations for best-first search that allow controlled use of memory, in: IJCAI-89 (1989) pp. 274–277.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Mahanti, A., Ghosh, S., Nau, D. et al. On the asymptotic performance of IDA*. Annals of Mathematics and Artificial Intelligence 20, 161–193 (1997). https://doi.org/10.1023/A:1018980327559
Issue Date:
DOI: https://doi.org/10.1023/A:1018980327559