Simulation Comparison Parameters (Last part)

(IV) RESOURCE UTILIZATION

The most common indicator used in practice is average resource utilization. This indicator provides good overview of the shop performance.

The average resource utilization can be defined as summation of resource utilization over all resources divided by total number of resources. Logically, a higher value of the indicator is a sign of better performance.

Average Resource Utilization = ∑ (Resource utilization across all resources) / No. of resources


(V) SERVICE LEVEL

This is one of the most common performance indicator used to compare the simulation results. In fact, this is one of the highly acceptable indicators to evaluate performance output of planning and scheduling activity.

Service level indicator is normally calculated in percentage.

Service level is defined as percentage of the orders completed within the due date criterion. This means, it gives the percentage of the orders those are completed with zero tardiness.

Service Level = (Total No. of Order – No. of orders with zero tardiness)/(Total Number of orders)

The above value multiplied by 100 gives you the percentage value of service level.


(VI) RESOURCE USAGE COST

There can be multiple derivatives for calculating production cost and related indicators. As at present, in Incoplan, only one cost parameter is defined, which is cost of resource usage, which in turn can provide a fraction of production cost.
This value can be defined as production run cost or resource usage cost.
The value can be defined as:

For all resources -
Resource usage cost = ∑ [Resource usage time*usage cost per unit time]

(VII) Q – RATIO

The Q-ratio is also one of the good indicators about your schedule. This is the ratio of work content to cycle time. As this indicator has direct reference to the work content, in a capital intensive industry, where the resource usage is critical, could provide a good guideline to understand the manufacturing performance.

Q Ratio = ∑ [Work content for all orders] / ∑ [Production cycle time of all orders]
There could be other different ways to calculate this value with respect to work content and cycle time.

Let us try to understand, what these calculating parameter means. We can again refer to the same problem, what has been defined earlier. In the following illustration, you can very well visualize meaning of work content and cycle time.


It is interesting to note that if the value of Q ratio is one [1], the sum of all work content is equal to the sum of production cycle times. This means, the unproductive time on resource is zero. Obviously, this is not possible in practice because of process constraints and priorities.
For the guideline, we can observe the following value pattern to identify comparatively better performance.

Simulation Comparison Parameters ( Part I )

OVERVIEW

It has been always a question of million dollars for the planning function to identify a good schedule considering demand, process constraint, capacity and stock. With advanced planning and scheduling tools it is possible to generate a good feasible schedule but also to create alternate solution for a given set of problem definition.

Once the multiple simulation results are achieved, it becomes even more complex to identify the best out of it for the given problem. We must understand that in any environment, process parameters and constraints are dynamic. Therefore, a good solution for a given set of parameters may not remain a good solution for new set of parameters.

Even further, the questions remains, how one can find a better solution amongst the given set of solution. It is interesting to learn and observe the different parameters, which enables us to compare these results on common platform and identify the ‘best-fit-case’ for current situation.

This knowledge papers is trying to identify few of the parameters, using those one can visualize and generate common platform to compare the simulation results for production scheduling problem.

The core idea here is to present a few core indicators, which are used commonly and provide basic information about these.

(I) MEAN CYCLE TIME

The cycle time is defined as the time units required to complete all operations of an order. One order may comprise of multiple operation with a particular sequence and constraints. The smallest completion cycle time for the order is equal to sum of time required for all operations.

In real life case, the operation may not be sequenced one after the other without a time gap. Therefore the actual cycle time required to complete the order is higher compare the ideal [smallest] cycle time required.

The mean cycle time is a good indicator for global level performance. This indicator provides value for the average cycle time of a group of orders selected / scheduled. The lower the value of the mean cycle time, the better the result is. This indicates the average time value, which an

order will take for completion with current set of orders.

(II) LATENESS OR TARDINESS

The same problem if viewed with respect to orders, the illustration can be as given below:
Here Cj = Completion Time and Dj = Due Date
Therefore Lateness Lj = (Cj – Dj)

a] Total Lateness = [L1 + L2 + L3 + ……………+ Ln]
b] Average Lateness = [L1 + L2 + L3 + ……………+ Ln] / n

Another criteria related to due date is Tardiness. Tardiness means number of jobs those are tardy in a given environment at a particular time, and it is not concerned with how much tardy a job actual is.
The tardiness of a job j is defined as: Tj = max {(Cj-dj), 0}

a] Total Tardiness = [T1 + T2 + T3 + ……………+ Tn]
b] Average Tardiness = [T1 + T2 + T3 + ……………+ Tn ] / n

(III) TOTAL WEIGHTED LATENESS / DELAY

In manufacturing environment with multiple manufacturing orders, the delay for each order has different value. For example, an order with low priority is delayed by good time units may be acceptable compare to a smaller delay for order with high priority.

Therefore, a new indicator can be defined for calculating lateness / delay, which is linked to the priority or importance [weight] of an order. This indicator normalizes the effect of lateness for priority. Consider again the same problem illustration:

Based on these parameters, we can calculate the indicator as given below.
Total Weighted Lateness / Delay = (P1*L1) + (P2*L2) + (P3*L3) + …….. + (Pn*Ln)

Similarly, average value of the weighted lateness / delay can be defined as:
Average Weighted Lateness / Delay = [(P1*L1) + (P2*L2) + (P3*L3) + …….. + (Pn*Ln)] / n

One can set standard indicator to compare these values and to find out whether the value of total weighted delay is good or not good.
For example: Standard Indicator = [average priority]*[k*Average production cycle time]
Here k is a multiplication factor, which is nothing but the acceptable deviation as percentage of average production cycle time. Here, the closer the value of average weighted lateness to the standard indicator, better the performance is.