ASCP notes - advanced supply chain planning lab PDF

Preview

Summary

advanced supply chain planning lab...

Description

INTRODUCTION OBJECTIVES OF THE LAB Connecting SCM practices, industry experience and knowledge about advanced IT tools to cope with advanced Supply Chain Planning tasks, in nowadays business scenario. Find the right interception between approaches planning, information techniques tools, and experience with specific industries. WHAT DOES ADVANCED SC PLANNING MEANS? -

-

MT Railways manufacturers 8 families of products, driven by available forecasts over a planning horizon of 10 months  PUSH PRODUCTION SYSTEM Production costs may experiment (small) variations during each month due to expected changes in components’ price and availability. Also selling prices may varyThe moment when you produce and the moment when you sell the items may be relevant for the profit of the company.

Production routings are well defined and stable, as well as resources’ capacity, which is limited. The trade channel could hold some inventories to cope with capacity limitations Setting up production for each family will trigger setup times (reducing available capacity) and extra costs, which have been quantified by manufacturing engineers. Main suppliers have capacity limitations as well, StellComp, for instance, has shared some available capacity plans, considering an aggregate supply lead time of 1 month and an aggregate coefficient of utilisation of 1 supply unit per each product.

Integrated planning exercise with push production system driven by forecasts and production cost is not constant. We do not have to deal with capacity planning (50% of the planning problem). Set up time reduces production capacity. How much capacity might be available for me in the future? This is not a linear problem: we have capacity limitation (model with linear equation), but if we have some I at t-1, then we produce something and we are going to sell something at t; we cannot sell more than the demand and more than was available + production, we need to introduce a minimum, or maximum and so it is not a linear problem. What I can sell is expressed by the minimum between the demand and what was available + what I was able to manufacture. To move this to a real SC level, we need few descriptive variables. DECISIONAL VARIABLES • •

P(i,t) = quantity produced of family i in month t SC (t) = aggregate supplied quantity from supplier SteelComp (arrives at time t) since supplier has a lead time we should order the quantity at time t-lead time 1

P(i,t)>0 and SC(t)>0 In order to have the problem well defined and easy the computational search because I am adding the boundaries SERVICE VARIABLE: they are useful to express the equations • • • • •

Sold (i,t)= actually sold quantity of family i in month t InvP-in (i,t)= inventory available for sale InvP-fin (i,t)= inventory position after sales are realized SET (i,t)= Boolean (0,1); 1 if the production of family is planned in month t  needed for setup, moves the problem to non-linear integer programming InvSC (t)= inventory position (end of period) of component SteelComp

RELEVANT DATA • • • •

D (i,t)= forecasted demand of family i in month t Price (i,t)= selling price SC-limit (t)= supply limitation at month t  not negative H (t)= hours available for manufacturing and setup

SOLUTION We can max profit or min costs: Min of costs include also opportunity costs 𝑚𝑎𝑥 ∑(𝑠𝑜𝑙𝑑𝑖,𝑡 ∗ 𝑝𝑟𝑖𝑐𝑒𝑖,𝑡 − 𝑃𝑖,𝑡 ∗ 𝐶𝑣𝑖,𝑡 − 𝑆𝐸𝑇𝑖,𝑡 ∗ 𝐶𝑠𝑒𝑡𝑢𝑝𝑖,𝑡 − 𝐼𝑛𝑣𝑃𝑓𝑖𝑛𝑖,𝑡 ∗ 𝐶𝐻𝑂𝐿𝐷 − 𝐼𝑛𝑣𝑆𝐶𝑡 ∗ 𝐶𝐻𝑂𝐿𝐷 𝑖,𝑡

𝑖,𝑡

𝑆𝐶

PRODUCTION CAPACITY CONSTRAINT: ∑ 𝑃𝑖,𝑡 ∗ 𝑇𝑖𝑚𝑒𝑖 + 𝑆𝐸𝑇𝑖,𝑡 ∗ 𝑇𝑆𝐸𝑇𝑈𝑃𝑖 ≤ 𝐻(𝑡) 𝑖

∀𝑡

RAW MATERIALS CONSTRAINT (availability constraint): We cannot produce more than the raw materials we have (cannot overcome the availability for raw materials in manufacturing) ∑ 𝑃𝑖,𝑡 ≤ 𝐼𝑛𝑣𝑆𝐶𝑡−1 𝑖

∀𝑡

what was available at the end of the previous period, is what we are going to drive out from the warehouse so we cannot produce more than what we have in the warehouse from the period before) COMPONENTS INVENTORY LEVEL: because we have to put some components in 𝐼𝑛𝑣𝑆𝐶𝑡 = 𝐼𝑛𝑣𝑆𝐶𝑡−1 − ∑ 𝑃𝑖,𝑡 + 𝑆𝐶(𝑡) 𝑖

Everything produced for every product

Because I need to put some components in, otherwise the warehouse Is empty

𝑆𝐶(𝑡) ≤ 𝑆𝐶𝑙𝑖𝑚𝑖𝑡 (𝑡 − 𝐿𝑇) → 𝑙𝑖𝑚𝑖𝑡𝑎𝑡𝑖𝑜𝑛 𝑡𝑜 𝑛𝑜𝑡 𝑣𝑖𝑜𝑙𝑎𝑡𝑒 𝑡ℎ𝑒 𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑟 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝑙𝑖𝑚𝑖𝑡𝑎𝑡𝑖𝑜𝑛

2

INVENTORY AVAILABLE FOR SALE: 𝐼𝑛𝑣𝑃. 𝑖𝑛(𝑖,𝑡) = 𝐼𝑛𝑣𝑃. 𝑓𝑖𝑛(𝑖,𝑡−1) + 𝑃𝑖,𝑡 SOLD QUANTITY: 𝑠𝑜𝑙𝑑(𝑖, 𝑡) = min(𝑑𝑖,𝑡 ; 𝐼𝑛𝑣𝑃. 𝑖𝑛𝑖,𝑡 ) INVENTORY AT THE END OF THE PERIOD 𝐼𝑛𝑣𝑃. 𝑓𝑖𝑛(𝑖,𝑡) = 𝐼𝑛𝑣𝑃. 𝑖𝑛(𝑖,𝑡) − 𝑠𝑜𝑙𝑑(𝑖,𝑡) Problems with this type of method (no linear programme regression) are: 1) Uncertainty (demand, replenishment LT, scarp rates) 2) Data accuracy and availability: data are wrong and inventory position in the company is wrong, or the inventory position in the trade channel in unknown; The 1st and the 2nd are connected because thing changes so the data may be not valid anymore. 3) Size of the problem: is really demanding (e.g. computational time can be an issue. Luxottica with millions of SKUs=35 thousands of SKUs) 4) Capacity planning: 50% of this course (not present in this case). A company can do a lot of stuff which is part of the problem as well; planning is about resources and activities. 5) Collaboration: some problems cannot be solved by yourselves, but you need to collaborate with your suppliers. it’s a way to simplify the problem. the solution is not in the algorithm, managing relationships, sharing risks, working on contracts instead of complicated them The course is about: 1. Understanding the practices and the topics we have been dealing with: inventory planning (what do I really know?) or cost/benefits assessment (new technology for increasing visibility in the supply chain), risk management tool. 2. Then there is something about the tools: understand what can be done with the tools available now in terms of core of the basics tools. (know already what are MRP or DRP) Data is the real important staff when we are going to planning, because without data we cannot plan anything. Loud platform and Big data analysis. Simulation is important since we can virtualize the supply chain. 3. Have some industry experience: apply some issues to some specific industries to understand the real problems there. planner’s job: use the world job because actually there is a problem with job actually. The planner’s job is one of those jobs which are going to impact the most by the appraisal of the IT capabilities and so to some extent some of the tasks, which are now assigned to a planner are easily made by next generation technologies; but at the same time the planning of a supply chain today is so bad, there is so much room to improve, that still this is a quite interesting position for those who can use the latest technologies and the best practices available. The point is up to that: what are the competences they are going to achieve thanks to this. A supply chain planning is the interception of this. 3

EXERCISE • •

• •

•

Sarah, a brilliant high school student, enrolls the Politecnico di Milano program in Management Engineering, and plans to attend the 5-year program to get a MSc degree. Tuition fees amount to €4,000/year, she spends €1,500/year for books and other recommended teaching materials; moreover, she bears extra-costs of about €400/month since she is living far from his family/home. A friend of her decided not to enroll to any university, and found a job immediately after the high school. He started from €1,100/month (13 months/year), with a 3% salary increase every two years. Since she wants to retire at the same age than his friend (he is retiring after 40 working years), Sarah plans to invest her first wage(s) in a financial program of the National Social Security institution, which has a cost of €20,000 that will allow her to retire 5 years before the standard. In this way, they will retire the same year. Assuming a cost of capital of 2%, please calculate what should be the value of Sarah's monthly salary (Assuming the same salary increase patters, 3% every two years) in order to break even.

Value of Sarah monthly salary in order to break even? 𝑑𝑖𝑠𝑐𝑜𝑢𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡:

1 (1 + 0.02)𝑡

In more to payback the after graduation first salary is 20159 /year, 33% more than her friend

OPERATIONS PLANNING- GROUND ZERO1

Supply Chain & Supply Chain Management

SUPPLY CHAIN= The entire network of firms who interact to turn raw materials into finished goods and services and to deliver them to end customers (Supply Chain Council)

Supply Chain Management is the integration of business processes from end user through original suppliers that provides products, services and information that add value customers. WHY SC MANAGEMENT? ▪ ▪ ▪ ▪ 1

The firm is not a stand alone entity but it is part of a network of interconnected firms Optimizing internal processes is no longer enough Need to manage processes going beyond the boundaries of the company Shift on competition: from “Company vs Company” to “Supply Chain vs Supply Chain” https://www.forbes.com/sites/onmarketing/2014/01/28/why-amazons-anticipatory-shipping-is-pure-genius/#c94527a4605e

4

 Network, processes, information and money managed as a whole in a dynamic environment SUPPLY CHAIN PLANNING

Ground zero Vocabulary: -

Tier upstream, echelon downstream Sell out: what you deliver to the customer, sell-in: what I’m pushing in the beginning of the chain, sellthrough: what is exchanged by an echelon in the middle of the chain

Existing reference models used in this field: -

SCOR model SCP matrix: companies use this model to map their own SCP processes

5

If we move to operations we have many frameworks such as APICS reference model. APICS (American Production and Inventory Control Society) is a reference framework provided by The Association for Operations Management. This literature model provides a standard framework to tailor on Operating Companies in scope. When you have complex problems you go down in sequence.

Very high level: strategic business plan (BP) connected to budgeting (ex. of FCA=60.000 cars per year) Going down, we have SOP, strongly focused on commercial families. Here, doesn’t care about how many Maserati of which type: it is a sales operation here. - MPS (it’s the master plan): unit analysis is the production family (something which is common from the manufacturing point of view), you work at a technical level (group SKUs). It’s connected to the Rough-CutCapacity- Planning. Deeper in details and shorter in horizon as you have to be more precise. MPS is crucial because it’s strongly connected with upstream planning and downstream planning. (ex. of FCA: a Maserati can be Quattro Ruote, sporty cars etc. We are at the product family from the technological point of view) - MRP: more detailed analysis, check, plan and purchase for the materials availability. Connected to the capacity- requirement-planning (CRP). 2 outputs: production orders and purchasing orders.  Times are changing and companies are looking at longer time horizons using the available technologies and go to a greater detail. In the MPS level we usually use a production family, in that case is not true anymore. It depends on the business. -

Target Strategic plan SOP MPS MRP Scheduling Performance measurement

Accomplish company’s vision, achieve financials Align demand and supply Minimize costs Synchronize flows Fulfil orders Learn, improve, motivate

Business object

Horizon/frequency

Time bucket

Product category

3-5 years/ once per year

Year

Commercial families Product families Purchasing categories/SKU SKU All of the above

1-2 years/quarterly 1 years/quarterly 3-12 months/ weeklymonthly Days-weeks/ daily-weekly All of the above, in the past

Quarter/year Month/ quarter Day/week Hour/day All of the above

6

DECOUPLING POINT Upstream pushing, downstream pulling.

If the customer is accepting always shorter delivery lead time, you have to redesign your products to be able to have an acceptable lead time, you need to reduce the variety in forecasts using modules. We can keep and stock few modules quickly assembled. If the customer is always accepting shorter and shorter delivery lead time, we need to move the modules closed to the final assemblies. Take care that this is a relative perception: your orders (as a manufacturer) could be another actor's forecast (as a retailer at the end of the supply chain). In a SC perspective you don’t consider a single company, if the retailers are working by forecast all the SC is ATS/MTS ; even though their forecast is my order. Very important strategic position connected with the service level. The more the risks we accept to carry, the faster we are toward the customers. Companies, for example Amazon with the prime service delivery, are trying to move the decoupling point closer to the customers (STS=ship to stock). This is a strategic decision strongly connected to the service level. Decoupling point: separates the part of the system managed "to order" from that the part managed "on forecasts".

MATERIALS REQUIREMENTS PLANNING It should be materials planning because material requirements planning is the name of the planning and the name of the solution. Production / purchase orders can be issued according to two different logics: 1) REQUIREMENTS BASED: 𝑟𝑒𝑝𝑙𝑒𝑛𝑖𝑠ℎ𝑚𝑒𝑛𝑡(𝑡) = 𝑓(𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦(𝑡), 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛(𝑡 ′ )) 𝑤ℎ𝑒𝑟𝑒 𝑡 ′ > 𝑡 Consumption could be both orders or forecast related to future consumption expectations Materials requirements = sequence of appointments: -

Starting from Final Assembly line schedule, Bill of Material, and on-hand inventory, requirements of parts and components are computed; 7

-

Requirements are lot-processed to take into account technical or economical constraints; Lots are “time-phased”, depending on specific production or purchasing Lead Times.

Information is transmitted upstream, to the first actor in the production chain, and it then "pushes" the flow of

materials. MRP doesn’t work because it considers: -

Infinite capacity lead time is an input (it should be an output) Large amounts of not accurate data MRP nervousness When we work with the company, there are a lot of inventories, because data are wrong, lead times are not true, and staff is not there.

A different way of doing things is INVENTORY MANAGEMENT: 2) STOCK MANAGEMENT: you just look at what you have now, this logic is much simpler but it requires an hard work in the SC structure and management 𝑟𝑒𝑝𝑙𝑒𝑛𝑖𝑠ℎ𝑚𝑒𝑛𝑡(𝑡) = 𝑓(𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦(𝑡)) Material requirements = replenishment, triggered by actual consumption - As available inventory reaches a certain level, a replenishment order is transmitted to the upstream stage. - Information goes upstream one stage at a time, thus “pulling” materials towards the end of the process, as required by the MPS. It’s smartest, you need forecasts (not in the operative view but in the engineering view): -

To select the proper inventory management model because there are thousands of inventory management model (EOQ, Poisson model etc) and we need to have an idea about how the demand is going to be in order to select the proper model

We need future expectations data: - To set the parameters of the model (reorder point= demand *LT) but we need the future demand and we use the past to have an expectation for the future. When to use stock management (pull)? CONDITIONS that are needed to use the stock management approach: -

supply has to be dependable (e.g. fuel always available)  you don’t know when you need to replenish the materials (e.g. fuel) has to be cheap otherwise we are not able to manage the gasoline, some fuel is always there. 8

-

the demand has to be stable: we are assuming that we are consuming it regularly.

You need to create the conditions to apply the stock management. Otherwise, you have to use an MRP because you want to purchase what you need. Therefore: - inventory keeping is not risky: o stable demand o items not perishable/subject to depreciation - inventory keeping is cheap: o cheap items DISADVANTAGES OF STOCK MANAGEMENT when I manage materials through echelon that are all inventory based, the demand signal, even if it was very steady, become discontinuous demand.  inventory management is one of the causes of the bullwhip effect: solution is batch size=1 Push (MRP tool)

Customer demand management

Pull (to order)

Decoupling point Push (on forecast)

Materials planning Pull (stock management) the bolt, in this case, is used to produce a component (engine) that is difficult to forecast. I start the productions just when the order arrives I can manage a bolt which is used to produce an engine that is pushed on forecasts (because it's quite easy to forecasts)

What's the use of stocks? Use of stocks: • To offer great service (" I want it and I want it now") • To reduce purchasing costs  speculative stocks • To optimize production and logistics costs EOQ, compromise between manufacturing and logistics costs • To dampen uncertainty (demand, production, transportation) safety stocks • To buffer/decouple operations phases • To save money (quantity discount, speculative stocks) • To allow for process transformation (e.g. whisky requires 12 years to be produced) • To show off • To transfer information to guide production activities with no information systems and to offer great service • To hide problems according to Japanese companies, use stocks to hide problems (you can’t synchronize, solve quality issues etc.) The same item can have multiple functions.

9

DISTRIBUTION PLANNING Decisions are: 1) Network 2) Transports 3) Inventories of finished product We have strategic long-term and short-term operative decisions:

Software: Sequential inventory management at the different echelons of the distribution chain DRP: during the 80’s DRP started to gain attention - Advanced DRP (also know as DRP II)

Retail works on forecast, collects data using DRP system; group them into the demand on the central distribution centre, aggregate into family and then use the information for the MPS. MRP and DRP are strongly connected and they work with the similar logic. One working in a synthetic way, the other in analytical way.

Optimize objective

Demand forecast

Lead times Nodes Synchronization Customized service level Costs taken into account

Sequential Max Service level, using stock at the local DC; doesn’t consider the whole distribution system Independent forecast at any echelon of the distribution chain Single level LT (mean and variance) None

DRP No optimization, just aggregates downstream requirements to link them with MPS and production Forecast of the final customer demand (independent), upstream stages are planned consequently Single level LT (mean and variance) Top-down, with no pegging

NO

NO

Inventory, stock out (implicit)

NO

DRP II APS TO optimize the whole SC

Forecast of the final customer demand (independent), upstream stages are planned consequently Full chain LT (overall mean and overall variance) Fully managed Manages also stock allocation to channels/ customers/stores Inventory, stock out, transportation

All advanced production scheduling (APS) systems, start from this integration and improve it: quick planning, handling more data, using actual LT instead of planned LT.

10

DESIGN SEAT CASE STUDY 1) Identify those signals which highlight relevant problems in the current Operations Management approach • • ...