Question 10.76: Using simulation to determine the number of dock bays in a r...

An existing warehouse is 50 years old and is to be abandoned as it requires considerable renovation to promote efficient warehousing practice. A warehouse is to be built adjacent to the existing site to handle all existing business. The question that has been asked is, “How many dock bays should be constructed?” The first thought is that waiting line analysis could be used to answer this question. Unfortunately, a surge of vehicles arrives in the middle of the morning and the middle of the afternoon. This prevents defining arrival rates in a manner acceptable for mathematical analysis. The best method of determining the number of dock bays appears to be simulation.

A time study is made of the arrival of trucks to obtain the data given in Tables 10.42 and 10.43. The time required to load and unload vehicles does not vary with the time of day. A time study of these activities results in the data given in Table 10.44 .

The truck spotting time is observed to be a constant of 0.1 hour. The existing warehouse has three dock bays. Considerable knowledge is available with respect to “typical” operations for three docks. Hence, the simulation model can be validated with three dock bays. The logic underlying the simulation model is as follows:

1. Initialize the model. Begin with three dock bays and add one bay each time it is reinitialized.
2. Generate a series of random numbers from 0 to 99.

3. Transform the random numbers to a series of truck interarrival times by relating random numbers to the cumulative frequencies in Tables 10.42 and 10.43. For example, in Table 10.43, random numbers 0 to 35 would result in an interarrival time of 0.125 hour, random numbers 36 to 76 would result in an interarrival time of 0.375 hour, and random numbers 77 to 99 would result in an interarrival time of 0.625 hour.
4. Generate a series of random numbers from 0 to 99 .
5. Transform the random numbers to a series of truck loading and unloading times by relating the random number to the cumulative frequencies in Table 10.44.
6. Assign trucks to dock bays, or if unavailable, to a queue. Unload the trucks, dispatch waiting trucks, and assign spotting time. Perform the truck loadings and unloadings for the entire day and maintain statistics.
7. Determine whether steady state is reached. If so, and six dock bays have been considered, print all statistics and terminate the model.
8. If steady state is reached, and less than six dock bays have been considered, return to step 1 .
9. If steady state is not reached, return to step 6 and simulate another day’s operation. By running this model, the types of data resulting from the simulation are

The data the model generated for three dock bays is seen to be consistent with what is experienced with the present operation. Therefore, the model is considered to be a valid representation of the truck being loaded and unloaded for various numbers of dock bays.

The data presented for four, five, and six dock bays can be evaluated by management, and a decision can be reached with respect to the number of dock bays to be built for the new facility.

The procedure described in the above example is typical of the simulation models applied in many areas of facilities planning. These models are often implemented using a specialized simulation language, rather than more general programming languages such as BASIC, FORTRAN, Pascal, C, or C++. In fact, there are several simulation languages that have been designed with material flow and facilities planning targeted as the principal application focus.

Among the simulation languages that are either designed for or well suited for facilities planning and simulation of material handling/manufacturing systems are ARENA, AutoMod, eM-Plant, Factory Explorer, GPSS/H (and SLX), GPSS World for Windows, MAST Simulation Environment, Promodel, Quest, Simscript II.5, Simul8, Visual SLAM, AweSim, Taylor ED, and Witness. Many of these softwares include animation and/or 3-D modeling capabilities. For information on some of the above packages, the reader may refer to listings periodically published by IIE Solutions, O R / M S Today, and other related publications.