Statistical aspects of simulation : An updated survey

Practical statistical techniques for the design and analysis of simulation experiments are presented. These techniques are relevant in both discrete and continuous, deterministic and stochastic simulation. To generalize and interpret the simulation output the analyst can use regression analysis. This analysis allows for interactions among factors. Actually the regression model provides either a first‐order or a second‐order approximation to the complicated simulation model. To decide which system variants (parameter combinations) should be simulated, the analyst may apply experimental design theory. This theory makes the exploration of the simulated system much more efficient and more thorough. In the preliminary phase of the simulation experimentation special screening designs can be used to investigate hundreds of factors in relatively few runs. In stochastic simulation additional problems arise. There are several approaches for determining how to initialize a simulation run and how long to continue that run. These approaches result in a confidence interval for the estimated response. Both steady‐state and transient behavior are examined. Special variance reduction techniques are briefly discussed; the use of common random numbers (identical seeds) is discussed in more detail.(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from another version of this item.)(This abstract was borrowed from (This abstract was borrowed from another version of this item.)