CHAPTER 1: IN COMMAND, BUT NOT IN CONTROL
What is supply network dynamics?
1. Supply: Operations Management
2. Network: Organization & Management
3. Dynamics: System dynamics simulation
A different world
The ideal world in 1950s and 1960s: integral control; to optimize the multiple stages; sub-optimization
of the whole was required → central control.
- Integrated firm that coordinates everything from a centralized and optimized decision-making
core.
- In reality, decision-making wasn’t all that central and clearly aimed at local interests, also
because of IT impediments → integrated firm with local and non-optimized decision-making.
- The rate of change in the environment has increased and therefore the need to respond
quickly raised. Increase in technological development, shorter product life cycles. Adapt to the
behaviour of the environment, which became more volatile.
Today’s supply chain reality: multiple, (semi-)independent organizational units, which local, ad-hoc
decision-making, who jointly serve final customers.
- The managerial response to the need for greater flexibility has been to fragment these
integrated chains into networks of semi-independent organizations.
- We need new structures and mechanisms that will assure coordination throughout the
network. Sharing information between the units is required.
- Manager response to the need for greater flexibility (= respond to changes) has been fragment
these integrated chains into networks of semi-independent organizations.
Examples of a new way
Example 1: Collaborative Planning
Independent companies collaborate in a network and they have to coordinate their activities. There
are no contracts or rules to do this. A way result in a successful collaboration, collaborative planning
can be executed. This will cause less upstream demand fluctuation.
Example 2: Sales & Operations Planning
Weekly collaborative S&OP meetings with representation of all supply network partners. All phases in
the network are monitored, forecast, and kept consistent and within control limits. It is important to
think about what is best for the company instead of what is the best for our own department.
Example 3: Collaborative KPIs
Create KPIs that are important for the company and customer so both are working to the same goal.
There will be a win-win situation.
Example 4: Operations planning with scenarios
Sales and operations jointly create differen scenarios of what could happen. The idea of that a ‘central
scenario’ will be executed. This is the scenario that is the balance of different scenarios so this is the
safest thing to do.
In all these examples, the hard work remains to turn around the vicious cycles of the soft factors into
virtuous ones. Transparency and trust between departments/companies is very important.
,A new mindset
We need a new perspective on management, one that starts from the notion of business dynamics:
Aspects of cognitive mindset Myopic management Business dynamic management
Time perspective Static Dynamic
Type of management data Flow rates; new customers, Accumulations: backlogs,
churn rate, actions this week installed base, workforce
View of causality One-directional Feedback causality
System boundary Local, independent Integral, independent
Approach to gaining Analysis: understanding little Synthesis: seeing the whole,
understanding pieces very well how the parts move together
Approach to analysis Back-of-envelope 5-time why
Aspects of cultural mindset Myopic management Business dynamic management
View of coordination Internal and external buyer- Organizational network, co-
suppliers relations; principal- production, collaboration
agent
View of collaboration Zero-sum game Win-win
View of IT IT as ‘them’, as outsourceable IT as integral part of workflow,
unit of business processes
Cost, volume Quality, reliability, customer
View on performance evaluation
satisfaction
Approach to risk and problems Problem fixing, ‘fire-fighting’ Problem prevention, high-
agility reliability
Course framework
1. Setting the stage (Chapters 1-3): a primer in supply chain dynamics, a primer in system
dynamics concepts and a primer in managerial decision-making;
2. Demand shock dynamics (Chapters 4-6): 3 real-world cases of demand shocks as experienced
by ASML, NXP and Interpolis;
3. Workload-Quality dynamics (Chapters 7-9): 3 real-world cases of product-development and
ramp-up dynamics as experienced by a semiconductor company, Airbus and KPN Telecom;
4. Collaboration dynamics (Chapter 10-12): 3 real-world cases of collaboration issues (so-called
relationship spirals) as experienced by Fokker, Philips Electronics and Atos Origin/KPN;
5. Moving forward (Chapter 13): Organisational change approach required to address supply
network dynamics, concluding remarks and Q&A session.
,CHAPTER 2: MODELS OF MAN
Stocks and flow modelling
Terminology used to distinguish between stocks and flows in different disciplines
Field Stocks Flows
Mathematics, physics and Integrals, states, state Derivatives, rates of change,
engineering variables, stocks flows
Chemistry Reactants, reaction products Reaction rates
Manufacturing Buffers, inventories Throughput
Economies Levels Rates
Accounting Stocks, balance sheet items Flows, cash flow or income
Biology, physiology Compartments Diffusion rates, flows
Medicine, epidemiology Prevalence, reservoirs Incidence, infection, morbidity
Example: a bathtub stock = water level in baththub, flow = inflow/outflow of water
Cloud icon: Any system dynamics is an open system, which means that there are stocks and flows
outside of this system.
Auxiliary: simple circle, denotes a piece of information that is used to drive one of the other flows, as
a good performance indicator.
Information link: arrow connecting the stocks, auxiliaries and rates.
There is a mathematical reasoning behind every model. For complex models, the calculation is very
difficult and sometimes impossible. Therefore, we use simulations.
Characteristics of stocks:
- Stocks characterize the state of the system (‘the snapshot test’)
- Stocks provide the basis for actions
- Stocks provide systems with inertia and memory
- Stocks are the sources of delays
- Stocks decouple rates of flow
- Stocks create disequilibrium dynamics
Four equivalent representations of stock and flow structure:
1. Hydraulic metaphor
2. Stock and flow diagram
3. Integral equation
4. Differential equation
, Dynamics of stocks & flow
Time step dt: how a continuous time model
is calculated in discrete time steps on a
digital computer.
Models of the real world should always
include feedback, of which there are only
two kinds: positive and negative feedback.
- Positive/reinforcing
- Negative/balancing
Negative feedback loops
Negative feedback is always goal-seeking. It will continue until it reaches its goal, form there on it will
not increase, decrease anymore. The behaviour will be corrected when it deviates from the goal.
- Implicit goal: 0 → exponential decay (rabbit population with rabbits deaths)
- Explicit goal: current value of stock equal to desired value of stock. Time it takes to adjust to
desired value is much longer than the adjustment delay.
It is about a gap between current value and a desired level. The model will work towards the desired
level until it reaches it. The gap between current and desired will get smaller.
Material delays
The delay is equal to the time adjustment. The greater the time adjustment, the longer it takes for the
current value to reach the desired value. It takes time to learn something in order to reach the goal.
Delays: special, common form of negative feedback. Behind delays, there are always stocks. The two
types are material delays and information delays.
Material delays: how the real world works.
- Most common form is 1st order delay (two stocks): Outflow = d * stockt
- 2nd order delay (three stocks), 3th order delay (four stocks), etc.
- Change of stock proportional to the change of the flow, flow is a certain percentage of the
stock.
- Higher-order delays are formed by cascading first-order delays together; middle stock will
become stable when the inflow equals outflow.
- S-shaped growth curve.
Information delays
Information delay: how information is processed by people and machines. It is used to represent
adaptive expectations, perceptions and other forms of human and machine information processing.
Over time, the perceived value will come closer to the currently reported value. Change in perceived
value goes slower than change in reported value. It takes some time before we adjust our opinion,
there is some noise over time. It can take a long time to figure out the correct value, ‘first impressions’
(= initially value = anchor) matter a lot. This mechanism also means that perceptions will always lag
behind reality.
Information delays smoothen out short-term (random)
variations in observations, which are caused by noise
Noise: random deviation, due to measurement errors.