These lecture notes are based on the subject Global Supply Chain Management of 2015-2016.

Lecture 1

Lecture 1 focuses mainly on:

• simulation
• Deterministic performance estimation

Both are approaches to to analyze and design a process.

There are different criteria to determine if a design is good or bad. A design is made in a flow diagram. These criteria are used to analyze the design. The main goal of analyzing a design, is to attain contiguous improvement.

Differences :

• Deterministic: easy and fast, but rough overview
• Simulation: long and complex, but gives a better and more specific outcome.

Simulation:

Simulation experiments many times, and tries to duplicate a real environment into its model. Mathematically, a real world situation is imaged. After wards, a conclusion is stated, and a company will decide if it will adapt to this conclusion.

Main advantages:

• It is used in difficult situations
• There is no investment needed
• Uncertainty is incorporated, so it is very functional in real life
• More solutions are shown

Main disadvantages:

• Analyzing and interpreting outcomes is difficult
• High computation time
• Difficult to make
• Value of the result is overestimated many times

Most of the large companies use this method.

To be able to understand the following topic, one must understand the following things:

• Throughput time
• Work-in-progress
• Bottleneck
• Throughput/departure rate
• Utilization & efficiency

Deterministic performance analysis:

There is a distinction between deterministic and stochastic determination time

• Deterministic: everything is known in advance
• Stochastic: things might change, not everything is known in advance

In order to know the deterministic throughput time, you must know the time between the point that the customer/product enters the system, and the costumer/product is ready. Always take the average time that is needed.

Design & effective capacity:

• Design: Theoretical maximum output
• Effective: capacity that can be expected given the circumstances

We distinct serial and parallel processing

• Serial: steps follow each other up
• Parallel: steps are done at the same time.

Parallel servers:

• One server can only serve one customer
• Other servers serve other customers simultaneous

You must be able to find the bottleneck in a system. Bottleneck = the part of a design that makes the rest of the design slower. It is similar to the statement: a design is a strong as its weakest link.

Departure rate = number of customers/products that leave the system per hour. Depends on the bottleneck of the system.

Utilization rate & efficiency:

• Utilization rate = total operating time/ total time = actual output/ design capacity
• Utilization efficiency = actual output/ effective capacity

Effective capacity = the time that could actually have been used.

Note: for the productive utilization rate, the set-up time is excluded.

Lecture 2

Work in progress (WIP) concept:

• The amount of products that have entered the system, but have not yet been completed.

To estimate the WIP, one uses Little's equation: Average WIP = arrival rate x deterministic throughput time

Another estimation is shown in the attachment. 

Only when the arrival process is the bottleneck, Litlle's equation can be used. So in order to use it: production rate > arrival rate

In order to make sure that a process which involves batches does not become the bottleneck, always make sure that you know the set-up time, and know the arrival rate. When the batch size is increased, the time that it takes to complete the process per product decreases.

Important notes:

• Know what simulation means

• Know how to identify the bottleneck

• Know how to make a batch size large enough, in order to prevent a process from becoming a bottleneck

Lecture questions, difficulties explained:

1.1A. Why are there no materials going from the Dutch Bank to the external transportation company? There is no flow of money to the external company, since it is directly transported from the Dutch Bank, to the other banks. As you can read because the trucks of the company stand on the parking lot of the Dutch Bank.

1.1B. Make sure that you find in the text what difficulties can be found, words like 'forecast' imply that mistakes can be made, so here can be decision problems.

2.2 Always make a graph where all information is shown. Always take the averages, so do not consider the standard deviations.

2.3 Utilization rate = total operating time/ total time = actual output/ design capacity

2.4 Arrival rate = amount of products/people entering the system per hour. For this question, it is 60/7. Throughput time = (3+8+11)/60

Lecture 3

Arrival rate= average number of arrived units per time-unit

Production rate = average number of produced units per time-unit

Service processes: Services are intangible, so can not be touched and also not be stored. It is different from the production process because for the production system tries to achieve as much production as possible. In a service system, it is about the quality of the service it can provide for the customer.

In services, the organization has to find the right balance, not too much capacity, since it will be very expensive and not too less capacity, since it will lose customers. So an organization has to make sure that hires sufficient employees or have enough technology.

Waiting-line models: Is a purely logistic problem, where you mathematically can find out how to effectively make sure that a queue will be as short as possible, or the system will be as profitable as possible. The performance can be measured by looking at the capacity and variability.

Different types of models are shown in the power point presentation.

During the service process, the service time per customer is uncertain, since almost nothing will be always equally fast. Customers themselves are also very different, some are impatient and will leave the queue.

We distinct the following customers:

• Balking: decides if he stands in line when he examines it.

• Reneging: customer leaves if it takes too long.

• Jockeying: customer constantly switches to the queue which seems the fastest to the.

We normally look at the arrival process as a Poisson process. This is explained in the slides.

2 queuing models: M/M/1 and M/D/1

Difference: M/M/1 is a negative exponential with regards to service time, while M/D/1 has a constant service time distribution.

For question 1: use M/M/1 because the service time distribution is negative exponential service time. Arrival rate = 60 per hour , the average production rate = 180 per hour.

Read the question carefully: it asks how long it takes for the customer before it can start drinking coffee. This means that it you also have to take the time it has to wait for the coffee to be made. So use formula Ws: 1/ 180-60 = 1/120 hours = 30 seconds.

For question 2: it takes exactly 15 second to make a cup of coffee, so use M/D/1. Arrival rate = 144 people per hour. Production rate = 240 per hour. It is about how many people wait in the queue, so use Lq 144^2/ 2*240(240-144) = 0.45

For question 4: First look if all the applicants can handle 8 people per hour, they can, since the slowest applicant, Martha, can handle 10 per hour. Ken can handle 12 and Eddie can handle 15. Now use Ls, to determine the average number of customers in the system, since it cost 4 Euro per hour to have a customer in the store.

For Martha: it means: 8/ 10-8 = 4 customers. 4*4 = 16 Euros. Total cost = 16+6= 22 Euro

When doing this for Ken it will cost the least.

To understand the relations between the formula's, look at the slides.

Lecture 4

Comparing both systems: M/M/1 will always have a larger amount of customers in the system than M/D/1. M/D/1 has a better performance than the M/M/1. The only difference is the service time, for M/D/1 it is constant, for M/M/1 it is uncertain. Uncertainty will degrade the value of a system.

Question 2 explained: use M/M/1 because the service time is uncertain. Arrival rate= 10, service time = 12. Use formula Ws, so 1/ (12-10)= 0.5, so it will take 30 minutes on average.

Question 5 explained: arrival rate = 10

For the exponential one, use M/M/1 with a service rate of 12. use Ls: 10/ (12-10) = 5

For the constant one, use M/D/1 with a service rate of 11. use Ls: 10*(2*11-10) / 2*11(11-10) = 120/22 = 5.45 cars. Difference = 0.45 in the advantage for the first one.

Question 4 explained: arrival rate = 60/7 per hour, service time = 20 per hour. Having more than 3 patients in the system: us Pn>k with k=3. This gives (60/7 / 20) ^4= 0.034

Question 3 explained: arrival rate = 10, service time = 12. Use 1- Pn>0

1- 10/12 = 1/6

Question 6 explained: arrival rate = 15, service rate = 20. use M/M/1 because of the negative exponential distribution. First find out what the chance is, that there is more than 1 customer in the system. Pn=1: (15/20)^2=….

now find out what the chance is that there are more than 0 customers in the system: Pn=0: (15/20)^1+….

To determine the probability that there is exactly 1 customer in the system = Pn>0 – Pn>1 = 0.1875.

Question 19 explained: important information: starting up system: 15 minutes. Making a booking costs approx. 10 minutes and inter-arrival time = 12 minutes. For every booking, she uses 2 minutes shorter than the inter-arrival time. She has to make up 15 minutes per batch in order to process everything. 15/2 = 7.5, this means she needs batches of 8 in order to make up the 15 minutes. Official calculations is in the slides. This is just a simpler method.

Question 20 explained: Make sure you make the bottleneck as small as possible. Process 4 takes the longest. Answer C will be the shortest one.

Question 21 explained: arrival rate = 12 per hour, service time = 60/7 per hour. To have 4 or more customers in the waiting line, this means there should be 5 or more in the system. Use Pn=5: (12/(60/7))^(4+1)=…

Question 22 explained: The M/D/1 model has a better performance because of the lower uncertainty. This means answer C is right. 

Lecture 5 - Guest lecture

Manufacturing simply means: Making stuff. 

The process Metrix shows how much a product is made, and how much it is standardized. This is in line with Organisational Structure.

Project production: Customer decides what a product will look like, most of the time large products, like ships or aeroplanes. A lot of products are put together in order to create one large product.

Jobbing: Is small-scale, also customized, so different products. E.g. gears that are in a machine, these always have to be slightly different. The perfect example will be for key manufacturers, they always have to make their product slightly different.

Batch production: Making more than 1 identical end-item. This means that batch production is more standardized. It will be the most efficient when the batches are as large as possible. Producing several products will mean that you have to change machinery etc. when switching to producing another product.

Mass production: Large volume of similar items. High formalization, trying to be as efficient as possible. Products like cars and mobile phones are examples of mass production.

Continuous production: Not discrete, only when measured one can define what it will cost etc. this goes for products like oil and sugar. The product also seldom changes, and is a discrete unit when measured.

There are different types of plant lay-outs:

• Fixed position lay-out: Like with the project production of a ship, all products are moved to the end product.

• Functional lay-out: associated with jobbing, Similar resources are grouped, because there is uncertainty about what exactly has to be made. This also goes for hospital patients, which are grouped in the kind of illness they have.

• Cellular lay-out: these have a wide spread of products are clustered together in order to be more efficient. There will be unique and standardized products.

• Product lay-out: There is a lot of similarity, this lay-out is most of the time associated with mass production.

Keep in mind that things will never be this discrete, this is just the theory, but things will never be exactly like this.

Customer order decoupling point (CODP): This means that you change a product for a specific customer order. First you can sell it to everyone, after the CODP you can only sell it to a specific customer.

Make-to-stock: A stock is produced in order for the customer to directly make them able to buy it. All work has already been completed, the customer only has to buy it.

Assemble-to-order: Like the subway sandwiches, products are already there, but will be assembled for the specific customer by order. There will be a longer waiting period than the make-to-stock.

Make-to-order: Several production steps are made when something is ordered. Even more time consuming.

Purchase-to-order: materials needed to make the product are only bought when the specific material is ordered. Most commonly happens with very expensive products.

Engineer-to-order: a specific product is only made for the first time when it is order. Happens with specific machines for large companies.

Only when the CODP has started, the customer notices what happens to a product. The more the CODP is close to the customer, the less the customer has to wait.

There will be upstream and downstream forces that will determine what strategy a company will use. Customers want their products as fast as possible, but a company does not want to have to much on stock. So, managing stock keeping units (SKUs) is very difficult, a company does not want too much stock.

Balancing line steps and determination of line efficiency are all in the book, and are explained on the slides.

Different types of lines are also explained in the lecture slides, where the effectiveness of the straight line and the U-shaped line are determined.

Lecture 6 

Lean manufacturing: only doing things that add value in the eyes of the customer. Explained by Japanese signs 'muda, mura and muri' in the slides.

One of the aspects is to produce the least waste possible. There are seven sources of waste, explained by the word 'TIMWOOD'. This is explained in the slides as well.

The Lean-Philosophy: never do more than is totally necessary. Also make sure that when you see other people doing more than needed, make them aware of this. Be as simple and efficient as possible. When problems are stated, make sure to attack them as fast and aggressive as possible. This will also make sure that your storage will be low, which is profitable.

The differences between the lean leadership and the traditional leadership are in the lecture slides.

This Japanese lean philosophy also gives the employees more involvement in the company. Also a lot of confidence and responsibility is given to the employees. It is shown that a lot of Japanese employees work for one company their whole life.

The lecture slides show which tools are very important and attractive in the lean philosophy. These aspects are:

• Value stream mapping: makes clear where the process is stuck, and where there is a lot of waiting time. This helps to decrease inventories and increase the flow.

• 5S: things will give a clear overview of what your resources are and how they can be used efficiently. This improves your workplace.

• 5Why's: asking through the obvious.

• Poka-Yoke: make sure that your product works every time.

• Reducing unevenness: Try to keep a constant and equal flow. Again: only produce what is needed and make waiting times as short as possible.

• Process vs. transfer batches: make smaller batches in order to make sure that all processors work, but have a small load. This reduces inventory. The Levelled scheduling is part of this approach.

• Just-in-time delivery

• Push vs. pull delivery: start at the last step, when inventory is needed, start working on it. Same as just-in-time. They use the pull strategy from Kanban. This is explained in the lecture slides. It is a simple approach.

• Visual control: in this way, problems are detected as soon as possible. If one of the people in the team finds a problem, it can stop the whole process.

• Kaizen: ongoing improvent. This means that you never know everything. Always try to improve yourself, because you are never the best possible.

There are different types of quality explained in the slides. It is important to distinguish these different types because for everyone, another quality is important. This means that quality is also subjective. The cost of quality is very hard to distinguish. In the slides, the cost of quality is divided by control cost and failure cost. Both will always occur, but failure costs are avoidable.

The average and the standard deviation of a process are very important. When you define these, you can see if a process is going according to plan. It is common to uses 3 standard deviations as a maximum to see if a process is in control. Also, R is measured to see if a process is going well. What you want of process is, that a process will be as close to the average as possible. This will make the process more easy to control, and more stable.

For the process capability, it is more important to be above the number that a company wants.

Six sigma is also an approach to have as least deviation as possible in a system. This is explained in the slides.

The 5 steps of total quality management are also explained very clearly in the slides.

Lecture 7

Sourcing is so important because many companies start to work on one only one thing. It easier for companies when it uses sourcing in order to work effectively on getting the best supply. A company needs supply in order to produce. Contribute to the lean philosophy, since it eliminates all unnecessary actions.

These suppliers do not only have to be the cheapest. The triple bottom line also has to be taken into account. The social and environmental responsibility of organisations becomes more and more important in order to create a good image.

A company can also integrate instead of outsource, for instance, Apple integrates its knowledge in all different countries by having departments there. But it outsources it product to China to be made.

Vertical integration: keeping things with the company

Strategic sourcing is meant for tight relations. So this means that companies have to think of how tight they want to have their relationship with their suppliers. Different relationships will occur with different suppliers.

The Sourcing Design Matrix in the slides shows what kind of relationship will be the best in different situations. The relationships are explained in the slides as well.

The bull whip effect is fully explained in the slides. The dangers, and why it will always happen will become very clear after looking at the slides.

Functional products are easy and everyday products that you as a customer will buy lots of times in your life. For innovative products, it is completely different. These are new products which demand very much attention from the organisation in order to make it in time to make profit.

The After Fisher chart will explain the differences between the products. The amount of certainty in a company will also form the strategy of an organisation, and it will also decide if the product will be functional or innovative.

Green sourcing is becoming very interesting for organisations nowadays, this is a way in which a company can stand out, and sell more products. It gives the company a good image as well, which will also generate more profit over time. Other profits are explained in the slides.

Lecture 8

Warehouses have multiple functions, these 4 functions are explained in the slides.

• Storage does not only help for what is explained in the slides, but also for shortening the lead time to the customer.

• Warehouses also help to change the mould for transportation

It is, however, very hard to choose a location, the important factors for choosing the location follow up the slide on the functions. To determine which factor should be weighted more, we use weights. This is subjective, but will help to come to a decision that seems most suitable in the eyes of the management.

For question 1, this subjectivity becomes very clear. In the graph of on slide number 17, the aspect of cost becomes clearer. The volume will show what will be the best city to open a new facility. This depends on the fixed costs and the variable costs. When the volume is low, it will be most handy to go for the city with the least fixed costs. When the volume should be very high, it will be most profitable to go for a city with low variable costs.

To calculate question 2:

Make the equations equal. 30000 +75x = 60000 + 45x

30x= 30000

x= 1000 units.

In order for Germany to be the better option, there should be 1001 units.

To use the formulas:

For x: multiply all the x coordinates with the volumes that belong to it, and divide it by the total volume.

This has to be done the same for y.

The numbers show where the warehouse can be placed the best.

Cross-docking: makes sure that the warehouse is very efficient in a way that the storage costs are very low.

On slide 24, the normal procedure in a warehouse is explained.

To design the warehouse, one must choose different aspects in the warehouse.

The I lay-out is meant in particular for cross-docking, since this will have a minimised storage.

The weighted distance: total distance travelled by all products.

Transportation is also very important, especially to determine which mode is suitable for the organisation. The mode means which kind of transportation, this can be a ship or an aeroplane. The criteria for choosing a mode are shown in the slides. Not only the transportation, but also the inventory costs have to be held in account. In the slides, it is shown how to compute the inventory costs.

For question 4: compute the inventory costs for the plane and the ship. This can be done by the following calculation for the plane: (1/365)x0.1x10000x250 and for the ship, (21/365)x0.1x10000x250, now calculate the other costs. You can conclude that the plane is cheaper.