MENDELSSOHN: USING MARKOV DECISION MODELS 



fishery — in the examples considered, some consis- 

 tency in the amount harvested is a desirable al- 

 ternative to high year-to-year fluctuations in the 

 harvest size. But this reduced the average per 

 period catch. Only in extreme situations, where 

 the cost of smoothing out the catch is greater than 

 the unit value of the catch, does any policy re- 

 sembling MSY become optimal. 



Finally, it is possible to obtain an understand- 

 ing of how robust the management measures are 

 to misspecifications of the underlying model. This 

 is important, since the model is only a guide to our 

 decision making, not the answer. In the models 

 considered, the "best" policies are robust in view of 

 this uncertainty. 



A question not examined is the assumption that 

 the population size is observed at the start of each 

 period. This too is usually costly, and inexact. Re- 

 cently, I and E. J. Sondik developed an efficient 

 algorithm that addresses the relative merits of 

 different sampling intervals for obtaining popula- 

 tion estimates.^ Together, all of these techniques 

 allow for an integrated, realistic approach to man- 

 agement under uncertainty. 



ACKNOWLEDGMENTS 



Debra Chow provided invaluable assistance in 

 programming the computer runs. David 

 Stoutemeyer of the University of Hawaii gave 

 much useful advice on maximizing the efficiency 

 of the optimization algorithm used. Lee Anderson, 

 George Fishman, and Adi Ben-Israel gave impor- 

 tant comments for improving an earlier version of 

 this paper. The paper also benefited greatly from 

 the comments of one of the referees and from C. 

 Walters who tempered some remarks in a previous 

 version. 



LITERATURE CITED 



ANDERSON, L. G. 



1977. The economics of fisheries management. The 

 Johns Hopkins Univ. Press, Baltimore, 214 p. 

 Bertsekas, D. p. 



1976. Dynamic programming and stochastic con- 

 trol Acad. Press, NY., 396 p. 



Clark, C. W. 



1976. Mathematical bioeconomics: The optimal manage- 



^Mendelssohn, R., and E. J. Sondik. 1979. The cost of in- 

 formation seeking in the optimal management of random re- 

 newable resources. SWFC Admin. Rep. H-79-12, 15 

 p. Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., 

 NOAA, Honolulu, HI 96812. 



ment of renewable resources. John Wiley and Sons, 



N.Y.,352p. 

 FOX, B. L. 



1973. Discretizing dynamic programs. J. Optim. Theory 



Appl. 11:228-234. 

 Fox, W. W.,JR. 



1970. An exponential surplus-yield model for optimizing 

 exploited fish populations. Trans. Am. Fish. Soc. 99:80- 

 88. 



1971. Random variability and parameter estimation for 

 the generalized production model. Fish. Bull., U.S. 

 69:569-580. 



1975. Fitting the generalized stock production model by 

 least-squares and equilibrium approximation. Fish. 

 Bull., U.S. 73:23-37. 



Hastings, N. a. J., and J. A. E. E. van Nunen. 



1977. The action elimination algorithm for Markov deci- 

 sion processes. In H. C. Tijms and J. Wessels (editors), 

 Markov decision theory, p. 161-170. Proceedings of the 

 Advanced Seminar on Markov Decision Theory held at 

 Amsterdam, the Netherlands, September 13-17, 1976. 

 Mathematical Centre Tract 93. Mathematisch Centrum, 

 Amsterdam. 



HINDERER, K. 



1978. On approximate solution of finite-stage dynamic 

 program. Universitat Karlsruhe, Fukultat fiir 

 Mathematik, Bericht 8, 42 p. 



Mathews, S. B. 



1967. The economic consequences of forecasting sockeye 

 salmon (Oncorhynchus nerka, Walbaum), runs to Bristol 

 Bay, Alaska: A computer simulation study of the potential 

 benefits to a salmon canning industry from accurate fore- 

 casts of the nms. Ph.D. Thesis, Univ. Washington, Seatr 

 tie, 225 p. 

 MENDELSSOHN, R., AND M. J. SOBEL. 



In press. Capital accumulation and the optimization of 

 renewable resource models. J. Econ. Theory. 

 PELLA, J. J., AND P. K. TOMLINSON. 



1969. A generalized stock production model. [In Engl, 

 and Span.] Inter-Am. Trop. Tuna Comm., Bull. 13:419- 

 496. 

 PORTEUS, E. L. 



1971 . Some bounds for discounted sequential decision pro- 

 cesses. Manage. Sci. 18:7-11. 

 REED, W.J. 



1974. A stochastic model for the economic management of 

 a renewable animal resource. Math. Biosci. 22:313-337. 

 KICKER, W. E. 



1958. Maximum sustained yields from fluctuating envi- 

 ronments and mixed stocks. J. Fish. Res. Board Can. 

 15:991-1006. 



Scarf, H. 



1959. Bayes solutions of the statistical inventory prob- 

 lem. Ann. Math. Stat. 30:490-508. 



SCHAEFER, M. B. 



1954. Some aspects of the dynamics of populations impor- 

 tant to the management of the commercial marine 

 fisheries. Inter-Am. Trop. Tuna Comm., Bull. 1:25-56. 

 VAN HEE, K. M. 



1977a. Adaptive control of specially structured Markov 

 chains. In M. Schal (editor), Dynamische optimierung, 

 p. 99-116. Bonner Mathematische Schriften 98, Bonn. 



1977b. Approximations in Bayesian controlled Markov 

 chains. In H. C. Tijms and J. Wessels (editors), Markov 



49 



