Note on Process Analysis

Process: A process is that which converts inputs to outputs. A process need not be a machine or a worker. For example, say we want to make tea. Then, fill water, heat water, place tea bag, and steep the tea, could be the four activities that are required to make tea. Put together they define the process of making tea. A slightly more involved process could be the process of admission to a movie theatre shown below.

Customers enter Queue for tickets

Customers arrive at ticket window

Credit verified and cashier returns card

Cashier keys info and swipes card

Customer tenders credit card

Cashier returns change

Customer tenders cash

Cashier asks for number of tickets and states costs and inquires payment method

Cashier hands tickets to Customers

Customers party enters hall

Security checks tickets and identifies seats

Resources: The things needed to carry out the process. These could be machines, labour, information, process recipes (how to make tea), material handling equipment, utilities (electricity and coal), etc. In the above figure the resources are cashier(s) and security person(s).

Process Flow Diagram: A process flow diagram shows activities, flows, and storage points where items wait for future transformation. It is a symbolic representation of the activities and how they interconnect with one another and is the starting point of any process analysis. The two symbols we shall use to draw a process flow chart are those of a square and a triangle. A square represents a process (a task/activity) and a triangle represents storage. For example see the sample process flow charts given in the class on Kristins Cookie, Donner or Manzana Insurance. Thus, the process diagram should indicate buffers and flows. Buffers stand for inventory (raw material (RM), work-inprogress (WIP), and finished goods (FG) inventory). The buffers in the movie admission process are customers in queue for tickets and customers queuing up to get hold of security to lead them to their seats (not shown in figure though). Flows stand for where the order goes from one process to another. In addition, process flow charts sometimes indicate information flows, decision nodes, who is doing what, line of customer interaction, and fail points. How do you make improvements to the process of admission to the movie theatre? You can ask questions like what is the average time it takes to hand-out tickets? What will be the impact of an additional cashier on customer waiting time? Should we have two different queues for cash payment and credit card payment? What are the implications of such a design? Should there be a security person (or persons) guiding customers to their seats or should we do away with it and allow customers to find the seats on their own or let customer sit where-ever they feel like. There are implications to these proposed changes.

Capacity and Cycle Time: The flow chart of a process is a static picture. To be able to analyze some of the dynamics of a process we need to define capacity and cycle time. 2

The capacity of a process is the maximum rate at which output can be created given an infinite supply of inputs and orders. Suppose you are observing the process at the exit of the system then the average inter-departure time is defined as the cycle time. If the process consists of only one task then the cycle time of this rather simple process is the processing time of the task.

For example consider the capacity of a process called heating water to make tea. Assume that the kettle can hold one gallon of water and that the time it takes to heat water to the appropriate temperature is five minutes. We say that the cycle time for this task is five minutes. The capacity of heating is 1 gallon per five minutes and in an hour we can do 12 cycles (60/5) and the resultant capacity of this task is 12 gallons per hour. Another example is of packing a computer. If it takes on the average 12 minutes to pack a computer, the capacity of the operation is 1 computer in 12 minutes or 5 computers per hour. The cycle time is 12 minutes. Remember that Capacity is a rate it is the maximum rate at which work can be done. It is quite important to understand that the cycle time is the average interdeparture time for a unit of the customer order. For example, the cycle time is five minutes per gallon and 12 minutes per computer in the two examples. If we place 50 cookies all at once in an oven to bake and the oven completes the process in 50 minutes, the cycle time is one minute per cookie. However, if in the last example the unit for analysis is 50 cookies then the cycle time is 50 minutes. Obviously, cycle time and capacity depend upon the unit of analysis. Usually, capacity is easier to determine but not always. The two concepts are indeed related. The minimum cycle time is defined to be the inverse of capacity. It has time as its unit of measurement. If demand rate is less than the capacity then the cycle time would be inverse of the demand. How do we measure cycle time? One method is to ensure an infinite supply of inputs and orders, then to stand at the output end of the process and record the time between orders (parts, cars, satisfied customers, etc.) exiting the process. For example, if we observe that a car comes off the end of an assembly line (as it does in most cases) roughly every one minute, then the capacity of the assembly line is 60 cars per hour or about 240,000 cars per year for two 8 hours shift operation and 250 working days per

year. There are issues of quality involved in these definitions, should we count the good parts or all parts? What is the correct definition? For our purposes we shall not count bad or defective parts as output.

Cycle Time: When we say that the cycle time of a machine is 10 minutes per part, we mean that if the machine were never starved for inputs it would produce on the average one good part every ten minutes. The notions of on the average and good part are very critical in this definition. Cycle time is the average time between parts exiting the process or task (if you were to stand and observe it). It is the inter-departure time from a process or a system. In these definitions, we have assumed that carrying out a process does not interfere with other processes. Unfortunately in many operations interference is rampant, i.e., if a person is assigned to perform two processes, the same person often cannot attend to these two tasks at the same time. When multiple constraining resources such as space, labour and machine are involved, the calculation of capacity is no trivial matter. The notion that subdivision of processes may be necessary for capacity analysis leads to the definitions of work area and system. In this note we shall also see how setup time, product mix and quality affect the capacity of a system.

Work Area: A grouping of similar processes is called a work area. For example, a collection of drilling machines represents the drilling work area at a Maruti plant. A collection of tellers comprises the customer banking work area at a Citibank branch. A collection of order takers comprises the customer order call center at Amazon or Dell, etc.

System: The collection of all processes viewed as a whole will be called a system.

Lead Time (also called flow time): The lead time is the time spent by a typical part, job, (order), or customer in the system. The total time spent in the system by a typical order is the system lead time. It is not generally equal to the sum of processing time of all the tasks in the system for one simple reason: there can be queues.

How do you measure lead time? Tag jobs or customers as they enter the system. Write down the time when they entered. See when they leave. Lead times for cars in an auto plant can vary from 12 to 24 hours often depending on how long they stay in the paint shop. The cycle time, however, can be around 1 minute.

Suppose there is a queue and a customer or order has to wait until served. This waiting time becomes part of the lead time (we will discuss this issue later in Capacity Analysis Under Uncertainty). It is clearly possible that each order going through a process could have different lead times. Thus lead time is not a unique number but a distribution. In this case to set a service guarantee we must decide what percent of the customer we want to satisfy in a make to order system. We call this percent the customer service level. In a make to stock or assemble to order system the service level will be determined by the lead time distribution as well as the level of finished goods or component inventories carried. We will cover these topics when we discuss inventory management.

Bottleneck: The resource with the smallest capacity in the system is called the bottleneck. The astute reader will note that we use the term resource and not activity in this definition. The reason is that the same resource may be required for carrying out two activities. Consider the example in which, one person verifies signatures on checks and then subsequently the same person verifies the balance in the customer's account. There are two activities, but the bottleneck is the person. Clearly having two people do the two jobs will increase the capacity. We shall assume unless stated that each process is carried out using a separate resource (or separate set of resources if there are more than one identical machines or workers attending to the same process). However we briefly describe how to analyze a process when the same resource (or set of identical resources) is used to carry out several activities. When the same resource is used to perform different activities do not compute the capacity of a activity but instead make a list of all resources. For each resource list the activities that have to be performed by that resource. Then determine the work that has to be done (in units of time). Finally, determine the capacity using the formula, capacity = time available (say 400 minutes in a day) divided by time required (work) per order (say 50 minutes).

As we shall see in our examples, the bottlenecks may shift depending on the demand pattern and the lot size. Improper scheduling also can create temporary and shifting bottlenecks.

System Capacity: The capacity of a system is the capacity of the bottleneck. The cycle time of the system is the cycle time of the bottleneck. Caution the capacity of the system is not the sum of the capacities of all resources but the capacity of the most limiting resource.

Idle Time: The idle time of a resource is the time during which work is not being done by a resource.

Transfer Time: This is the time during which the order is being moved for one work area to another. Typically no work is being done and no value is being added. Littles Law Rate in Process Rate out

Three operational measures of process performance: Lead Time (Flow Time) T (hours) Throughput rate R (units/hour) Inventory I (units)

Definitions: Flow rate (or throughput rate) is the number of units that flow through a specific point of the process per unit time. Flow time is the total time spent within process boundaries. Inventory is the number of units present within process boundaries at time t.

Littles Law: Average Inventory = Average Throughput Rate * Average Lead Time

Example (customer flow): A restaurant processes, on average, 1500 customers per 15-hour work day. At any point in time, there are, on average, 75 customers in the restaurant. Given: Throughput rate = R = 1500/day or 100 customers / hour Average inventory = I = 75 customers Derived: Time = I/R = 75/100 or average customer spends hour in the restaurant. Another example (job flow): A branch office of an insurance company processes 10,000 claims per year. Average processing time is 3 weeks. Assume that the office works 50 weeks per year. The process is a branch of the insurance company and the flow unit is a claim. Given: Throughput rate R = 10,000 claims/year. Average flow time T = 3/50 year. Derived: Inventory (average) I = R * T = 10,000 * 3/50 = 600 claims.

Lead time measurement, T: Direct measurement: 1. Observe the process over a specified, extended period of time. 2. Take a random sample of flow units over the specified period. 3. For each flow unit in the sample, measure its flow time from entry to exit. 4. Compute the average of flow times measured.

Indirect measurement: T = I/R.

Key managerial levers for lead time: 1. Decrease waiting time (to be covered later). 2. Decrease theoretical flow time a. Reduce the work content of critical path activities i. Eliminate non-value-adding aspects of the activity (work smarter). ii. Increase the speed at which the activity is performed (work harder). iii. Reduce the number of repeat activities (do it right the first time). iv. Change the product mix. b. Move some work content off the critical path.

Flow rate measurement, R: Direct measurement: 1. Identify a particular entry and exit point in the process. 2. Observe the process over a given, extended period of time. 3. Measure the number of flow units that pass through a selected point over the selected period of time. 3. Compute the average number of flow units per unit of time. Key Managerial Levers for Managing Flow Rate: 1. Manage supply and demand to increase the throughput. a. Have reliable suppliers; produce better forecasts of demand. 2. Decrease resource idleness to increase process capacity. a. Synchronize flows within the process to reduce starvation. b. Set appropriate size buffers to reduce blockage. 3. Increase the net availability of resources to increase process capacity. a. Improve maintenance policies, perform preventative maintenance outside periods of scheduled availability, institute effective problem solving measures that reduce frequency and duration of breakdowns.

b. Institute motivational programs and incentives to reduce absenteeism, increase employee morale. c. Reduce the frequency of or time required for setups or changeovers for a given product mix or change the product mix. 4. Increase the theoretical capacity. a. Decrease unit load on the bottleneck resource pool. i. Work faster, work smarter, do it right the first time, change product mix. ii. Subcontract or outsource. iii. Invest in flexible resources. b. Increase the load batch of resources in the bottleneck resource pool (increase scale of resource). c. Increase the number of units in the bottleneck resource pool (increase scale of process).
d.

Increase scheduled availability of the bottleneck resource pool (work longer).

Summary Definitions: Capacity = the maximum rate of output of a process, measured in units of output per unit of time. Capacity Utilization: Capacity utilization is a measure of how much output was actually achieved relative to capacity. If the capacity of a process is 500 units per day and on a given day 480 units are produced, then on that day capacity utilization was 96%. Bottleneck = the production resources that limits the capacity of the overall process (this is the step with the lowest throughput rate (capacity) or longest cycle time). Cycle time = average time between completion of successive units. It is directly related to output rate. (A throughput rate of 4 units per hour has a cycle time of 15 minutes). Lead Time (Flow Time): A span of time required to make a product or a service. It is the length of time spent in the process. Lead time refers to how long that unit takes between entering and leaving the process, including any queuing time.

Batch = number of units of a particular product type that is produced before beginning production of another product type.

For any process analysis following guidelines may be useful: 1. Define the processdetermine the activities and flow of information and goods. Also determine where inventory is kept in the process. This effort can be recorded in a process flow diagram. Determine the capacity of each resource. This may depend on the batch size (because if there is a setup time then larger batches means higher capacity for that resource). Given the batch size one can now determine the highest utilized resource. This will determine the capacity of the system. Example I In this example, parts are produced every five minutes. The cycle time is five minutes per part for the system since Machine A is the bottleneck operation, the flow time is seven minutes in the system (timed from the start of the processing in machine A) and the capacity is 12 parts per hour. This system is not balanced because the two machines A and B have different cycle times. It is indeed rare to see a perfectly balanced operating system. This is why the notion of a bottleneck is useful. If the system is fully utilized, then machine B is idle for three minutes in every five minutes, and we say that machine Bs capacity utilization is 2/5 or 40%.

Machine A Process Time = 5

Machine B

Process Time = 2

Finished Goods

10

Example II

Machine A

Machine B Finished Goods

Machine A Machine Process Time = 5

Process Time = 2

Work Area Cycle Time = 2.5

In this example, the cycle time for the work area with two machines of type A is 2.5 minutes. This work area is the bottleneck (note that the whole area is the bottleneck not just one of the two machines denoted as A). The capacity of the system is 60/2.5 = 24 parts per hour. (Conversely, the capacity of each machine A is 12 parts per hour. So the capacity of the work area is 24 parts/hour. The cycle time = 60/24 = 2.5 minutes.) If the two machines A actually produced 20 parts per hour yesterday, we will say that the capacity utilization was 20/24 = 83.33% yesterday. Computing Utilization in Example II We are given that the cycle time of the bottleneck is 2.5 minutes. In the month of January 2006 we find that the system produced 3100 units. What was the capacity utilization of the bottleneck? Assume that there are 20 working days in the month and 7 working hours per day. Basic Approach: Reduce everything to common units e.g., time. Time available per machine = 20 x 7 x 60 = 8400 minutes Process A: Time available = 2 machines x 8400 = 16,800 minutes. Time used = 3100 units = 5 x 3100 units = 15,500 minutes Capacity utilization = Time used/Time Available = 15,500/16,800 = 92.3%

11

Process B: Time available = 8,400 minutes Time used to produce the 3100 units = 3100 x 2 = 6,200 minutes Capacity utilization = 6,200/8,400 = 73.8% Bread Manufacturing The bakery operates two parallel lines, each equipped with a mixer, a proofer, and an oven. In addition, the bakery operates a single packaging line which is shared by the two bread making lines. A process flow diagram is shown below for the bakery.

Raw material

Work in Process

Figure: Bread-making Process

The processing time for the three steps is given below. The batch size is 100 loaves. Mix = hour / 100 loaves

Proof = hour / 100 loaves Bake = 1 hour / 100 loaves

1. What is the bottleneck in the bread-making process? 2. What is the cycle time of the system?

12

Raw material

Bread-making Process

Work in Process

Packing

Finished goods

Figure: The entire Process

3. If the packing line operates at 1 hour / 100 loaves, then what is the capacity of the system (in number of 100-loaves per day, assuming 8 hour-day)? 4. What is the cycle time? 5. In the beginning of the day, when the system is empty, then how long will it take for the first batch (of 100 loaves) to become finished goods? This is the throughput time.

(b) If the mixer, proofer and oven were not set up as two distinct production lines and products could flow from each mixer to either proofer and each proofer to either oven as shown below, then what is the capacity of the bread-making process? The processing time have not changed.

Raw material

Work in Process

13

From the work-in-process (WIP) it next goes to the packing line, then what is the capacity of the entire system?

Muffin Manufacturing

The dough and filling for the muffin are prepared separately. The parallel processes of making dough and mixing filling are dependent, both must be completed before the muffin can be filled, folded and baked.

Mixing the filling for a batch of 50 muffin is relatively quick, taking only 10 minutes. However, preparing the dough involves three steps: mixing, proofing and rolling & cutting. Proofing, with a process time of 15 minutes per 50 muffin, is the slowest step in the process of preparing dough, with mix and roll & cut taking 5 minutes per 50 muffin.

Although a new batch of filling could be prepared every 10 minutes, the batch of dough necessary to begin the fill & fold task can only be available at 15 minute intervals. The fill & fold task takes 5 minutes per 50 muffin, the bake task takes 20 minutes per 50 muffin and the pack task takes 10 minutes per 50 muffin.

14

1. Complete the process flow diagram and label processing time for each step. 2. What is the cycle time of the dough making and mixing the filling process? 3. What is the cycle time of the entire process? 4. What is the 8 hour daily capacity of muffins?

We visit an office that processes insurance claims. The manager asserts that it takes just two days on the average to process a claim. We examine the past records and see that on the average 200 claims are processed to completion every week. A quick estimate of the pending claims in the office reveals that there are 1500 claims waiting to be processed at different desks (stages of processing) in the office. 1. What is the throughput rate in this example? 2. Can we say anything at all about the capacity of the office for processing claims? 3. What is a good estimate for the lead time in the office?

15