An Investigation of Sequential Sampling Method for Crossdocking Simulation Output Variance Reduction

Abstract

This paper investigates the reduction of variance associated with a simulation output performance measure, using the Sequential Sampling method while applying minimum simulation replications, for a class of JIT (Just in Time) warehousing system called crossdocking. We initially used the Sequential Sampling method to attain a desired 95% confidence interval half width of ± 0.5 for our chosen performance measure (Total usage cost, given the mean maximum level of £157,000 and a mean minimum level of £149,000). From our results, we achieved a 95% confidence interval half width of ± 2.8 for our chosen performance measure (Total usage cost, with an average mean value of £115,000). However, the Sequential Sampling method requires a huge number of simulation replications to reduce variance for our simulation output value to the target level. Arena® (version 11) simulation software was used to conduct this study.

Investigating the Effectiveness of Variance Reduction Techniques in Manufacturing, Call Center and Cross-Docking Discrete Event Simulation Models

“…This variance is usually that which is associated with the performance measure under consideration. For example, (Adewunmi et al 2008), investigated the use of the Sequential Sampling Method (Law and Kelton 2000) to achieve a target variance reduction for a selected simulation output performance measure. Results from experimentation indicate that this technique for reducing variance requires a huge number of simulation runs to achieve any success for this particular simulation model.…”

“…There are various alternatives for dealing with the problem of improving the accuracy of simulation experimental results. It is possible to increase the number of replications as a solution approach, but the required number of replications to achieve a desired precision is unknown in advance (Hoad et al 2009) , (Adewunmi et al 2008). Another solution is to exploit the source of the inherent randomness which characterizes simulation models in order to achieve the goal of improved simulation results.…”

Optimisation of a Crossdocking Distribution Centre Simulation Model

“…For example, manual order picking operators can have different skill levels and familiarity with picking certain types of orders, while automated order picking machines failure are sometimes random occurrences. These arbitrary events amongst others can influence the overall volume of orders fulfilled through the Crossdocking distribution centre [Adewunmi et al 2008]. In such a situation, it becomes important for the achievement of a smooth Crossdocking operation, to pay particular attention to the order picking process within the Crossdocking distribution system [Li et al 2004].…”