Wei, Dai, & Schonfeld ABSTRACT A numerical method has been developed for estimating delays on congested waterways represented by series of G/G/1 queues, ie, with generally distributed arrival and service times and one chamber per lock. It is based on a metamodelling approach which develops simple formulas to approximate the results of simulation models. The functional form of the metariiodels is derived from queueing theory while their coefficients are statistically estimated from simulation results. The algorithm scans along a waterway and sequentiaUy estimates at each lock the arrival distributions, departure distributions, and delays. It can be applied to systems with two-way traffic through common bi-directional servers as well as to one-way traffic systems. Computational results are presented in this paper to illustrate the speed and convergence properties of the algorithm and to investigate some of its variants. The algorithm works satisfactorily and flexibly with different convergence criteria and scanning processes. For an illustrative 20-lock system, parameter estimates converge within five iterations and less than three seconds of CPU time to differences lower than 0.1 percent between successive iterations. The computation time is found to increase only linearly with the number of locks in the system, thus allowing the analysis of very large systems of interdependent queues.