Samenvatting Designing and Managing the Supply Chain PDF

Preview

Summary

Supply Chain Management Book Summary! Chapter 1 What is SCM? a set of approaches utilized to efficiently integrate suppliers, manufacturers, warehouses and stores, so that merchandise is produced and distributed at the right quantities, to the right locations and at the right time, in order to minim...

Description

Supply Chain Management Book Summary! Chapter 1 ? – a set of approaches utilized to efficiently integrate suppliers, manufacturers, warehouses and stores, so that merchandise is produced and distributed at the right quantities, to the right locations and at the right time, in order to minimize systemwide costs while satisfying service level requirements. -

The definition leads to several observations: o Every facility hthat has an impact on costs and plays a role in making the product to customer requirements must be considered o Objective = to be efficient and cost effective across the entire system! o Encompass the firm’s activities at many levels, from strategic to tactical to operational.

What makes SCM difficult? -

-

Supply chain ation! Directly affected by another chain i.e. development chain (set of activities associated with new product introduction) Challenging to design such that total systemwide costs are minimized and service levels are maintained.

-

-

-

-

in every supply chain.

the set of activities and . o Specifically includes decisions such as product internally and what to buy from suppliers ( early supplier strategic

what to make decisions, supplier

! However, in most orgs, diff managers are responsible for the diff activities that are part of these chains. Unless carefully addressed, the typical impact of this organizational structure is a misalignment of product design and supply chain strategies. to make matters worse, in many orgs, additional chains intersect with both the development and supply chains (e.g reverse logistics chain)  INSERT FIG 1.2

Global Optimization ( 1) the supply chain is a . 2) Different facilities in the supply chain frequently have different and c

a. Thus suppliers goals are in direct conflict with the manufacturer’s desire for flexibility. (as the like to produe in large quantities) 3) The supply chain is a that evolves over time (not only do customer demand and supplier capabilities change over time, but also supply chain relationships. 4) System - (the planning process needs to account for demand and cost parameters varying over time due to the impact of seasonal fluctuations, trends, advertising and promotions etc) Chapter 2 – Inventory Management and Risk Pooling -

Inventory can manifest itself in several forms: o Raw material inventory o WIP inventory o Finished product inventory. (each of the above needs its own inventory control mechanism!)

-

? o Unexpected . Uncertainty has increase due to  (hence lack of historical data)  Presence of products in the marketplace o Presence in many situations of a in the quantity and quality of supply, supplier costs and delivery times. o o offered by tpt companies – encourages firms to tpt large qtys and hence hold large inventories! Similarly, incentives provided by manufacturers to distributors and retailers motivate buyers to purchase large qtys during manufacturer’s promotional periods and hence lead to high inventory costs. (2 critical issues = what to order and when to order)

-

To decide on an effective inventory policy, must take many characteristics of supply chain into account: Customer demand Replenishment lead time The number of diff products being considered Length of planning horizon Costs (including order costs and inventory holding costs) a. Order costs = Cost of product, and transportation costs b. Inventory holding costs = property taxes, maintenance costs, obsolescence cost etc. Service level requirements.

1) 2) 3) 4) 5)

6)

Single stage inventory model -

The Economic Lot size model

o Illustrates the trade off btw ordering costs and storage costs. Assumes the following:  Demand is constant (D)  Order qtys are fixed (Q)  Fixed cost (Setup cost) (K) is incurred each time the warehouse orders  Inventory carrying cost (h) (holding cost) is accrued per unit  Lead time = 0  Initial inventory = 0  Planning horizon is long o Extremely simplified version of the real inventory system. The assumption of known fixed demand over a long horizon is unrealistic. o It is easy to see that in an optimal policy for the model, orders should be received at the warehouse precisely when the inventory drops to zero. Called the zero inventory ordering property, which can be observed by considering a policy where orders are placed and received when the inventory level is not 0. o Total inventory cost = K + hTQ/2 o Average inventory level = Q/2 o  FIG 2.3 o HENCE, QUANTITY THAT MINIMIZES COST FUNCTION: √𝟐𝑲𝑫

-

 Q* = 𝟐 o 2 IMPORTANT INSIGHTS: 1. An optimal policy balances the inventory holding cost per unit time with the setup costs per unity time! (KD/Q = hQ/2) 2. Total inventory cost is insensitive to order quantities i.e. changes in order quantities have a relatively small impact on annual setup costs and inventory holding costs! (Sensitivity analysis in Table 2.1) The effect of demand uncertainty: VERY IMPORANT PRINCIPLES! o The forecast is ALWAYS WRONG o The LONGER the forecast horizon, the WORSE the forecast! o Aggregate forecasts are MORE ACCURATE! (while it is difficult to predict customer demand for individual SKUs, much easier to predict demand across all SKUs within one product family! This is an example of RISK POOLING!)

SINGLE PERIOD MODELS -

-

The optimal order quantity is not necessarily equal to forecast/average demand. It depends on the relationship between MP and MC. Note: FIXED COST has no impact on the production quantity, only on the decision on whether to produce or not. As order qty increases, average profit typically increases. Parabola shape (average profit curve) Risk-reward tradeoff.

Initial inventory -

having initial inventory helps to avoid paying the fixed cost for ordering goods. (s,S Policy) i.e. min max policy. s = reorder point, S = order up to level/max.

Multiple order opportunities. -

-

-

-

consider the fact where a distributor faces random demand for a product and meets that demand with product ordered from a manufacturer. (however, cannot instantaeneously satisfy demand due to fixed lead time) 3 reasons why distributor holds inventory: 1) To satisfy occurring demand during lead time 2) To protect against uncertainty in demand 3) To balance annual inventory holding costs and annual fixed order costs. 2 types of distributory policies: o Continuous review policy – where the policy is continuously reviewed and an order is placed when the inventory reaches a particular level.  The following assumptions are made:  daily demand is random and follows a normal dist.  Each time distributer places an order. FC K is charged  Inventory holding cost is charged per item.  Inventory level is continuously reviewed, and if an order is placed, the order arrives after the appropriate lead time  If customer order arrives when there is no inventory on hand to fill the order, inventory is lost.  Dist. Specifies a required service level. Service level = the probability of NO STOCK OUT during LEAD TIMES. For formulas, read page 43.  Reorder level = L * AVG + z * STD * √𝐿 2𝐾∗𝐴𝑉𝐺  Optimal Q* = √ ℎ 

Average inventory level =



INSERT FIG 2.9

𝑄 2

+ z * STD * √𝐿

If the lead times are variable, Reodrer point R = AVG * AVGL + z√(𝐴𝑉𝐺𝐿 ∗ 𝑆𝑇𝐷2 + 𝐴𝑉𝐺 2 ∗ 𝑆𝑇𝐷𝐿2 ) o Where z√(𝐴𝑉𝐺𝐿 ∗ 𝑆𝑇𝐷2 + 𝐴𝑉𝐺 2 ∗ 𝑆𝑇𝐷𝐿2 ) = 𝑠𝑎𝑓𝑒𝑡𝑦 𝑠𝑡𝑜𝑐𝑘! Optimal Q* is the same! o Periodic review policy – where the inventory is reviewed at regular intervals, and an appropriate quantity is ordered after each review.  Unfortunately, the (Q,R) policy cant be directly implemented since the inventory level may fall below the reorder point when the warehouse places an order.  Base stock level – the warehouse determines a target inventory level, the base stock level, and each review period, the inventory position is reviewed and the warehouse orders enough to raise the inventory to this level.  What is an effective base stock level?



 

-

The level should be enough to protect the warehouse from shortages until the next order arrives! Since the next order arrives after a period of r+L days, the current order should be enough to cover demand during a period of r+L days! (logical) Insert Fig 2.10. Expected level of inventory after receiving order: = r * AVG + z * STD * √(𝑟 + 𝐿) NOTE THAT THE 2nd PART OF THE EQUATION IS ALWAYS EQUAL TO THE SAFETY STOCK!

SERVICE LEVEL OPTIMIZATION o So far, we have assumed the the obj of inventory optimization is the det the optimal inventory policy given a specific service level target. However, how should the facility determine the appropriate level of service? Sometimes, this is determined by the downstream customer. In other cases, facility has the flexibility to choose the appropriate level of service. However, the Trade off is: – ! Thus, we give a target service level across all products and determine service level for each SKU such as to maximize expected profit. All else being equal, service level is higher for pdts with : ! ( Insert FIG 2.11!!)

-

Risk Pooling o Suggests that demand variability is reduced if one aggregates demand across locations. As we do so, it is likely that high demand from one area will be offset by low demand from another area. To understand this, must understand the coefficient of variation and the standard deviation!  Difference: Std Deviation = ABSOLUTE variability of customer dd Coff var = variability relative to avg dd!

 3 very critical points about risk pooling: 1) Centralizing inventory reduces both safety stock and average inventory a. In a centralized dist system, whenever demand from one market area is high than avg and another is lower that avg, items in the warehouse can be reallocated to the other! This is not possible in a decentralized dist system where diff warehouses serve diff markets. b. The higher the coefficient of variation, the greater the benefit obtained from centralized systems i.e. greater benefit from risk pooling. Avg

inventory has 2 components: weekly demand and safety stock (i.e. std deviation of weekly demand) since reduction in avg inventory is achieved mainly due to a reduction in ss, the higher the coeff of variation, the larger the impact of ss on inventory reduction! c. The benefits of risk pooling depend on the behaviour of demand from one market relative to another. Dd is positively correlated that when one is greater than avg, the other is also greater than avg! -

Centralized vs decentralized systems.

Safety stock Service level Overhead Customer Lead Time

Centralized Less ss Higher Lower (Due to EOS) Higher

Transportation costs

Depends

-

Decentralized More ss Lower Higher Lower (Due to warehouse closer hence shorter response time depends

Managing inventory in the supply chain o Consider a multifacitlity serial supply chain that belongs to single firm. We assume that:  Inventory decisions are made by a single decision maker whose obj is to minimize systemwide costs.  Decision maker has access to inventory information.  Base this on an echelon inventory policy.  Each stage/level is refered to an echelon. Thus. The echelon at any stage of the level of the system is equal to the inventory on hand at the echelon + all the downstream inventory! (downstream means closer to the customer!)  INSER FIG 2.13  Formulas for echelon supply chain:  R = L * AVG + z * STD * √𝐿  Q* is the same always!  R = 𝐿𝑒 * AVG + z * STD * √𝐿𝑒

Practical Issues – TOP STRATEGIES! 1) 2) 3) 4)

Perform periodic inventory review Provide tight management of usage rates, LT and safety stock Reduce safety stock levels Introduce/enhance cycle counting practice – replaces the annual physical inventory count by a system where part of the inventory is counted every day, and each item is counted several times per year 5) Follow ABC Approach 6) Shift more inventory ownership to suppliers 7) Follow quantitative approaches – inventory turnover ratio! ( Sales/Avg Inventory)

Forecasting – Remember the 3 things about forecasts: 1) the forecast is always wrong 2) the longer the forecast horizon, the worse the forecast 3) aggregate forecasts are more accurate! 4 ways to determine forecasts – judgement methods (Delphi, panel of experts), market research methods, time series methods (moving average, exponential smoothin, intrapolation) and causal methods.

Chapter 3 – Network Planning -

-

Network planning- the process by which the firm structures and manages the supply chain in order to o find the right balance btw inventory, tpt and manufacturing costs o match demand and ss o utilize resources effectively. Considering that this is a complex and hierarchical process, requires 3 steps: **  

1) Network Design (strategic decision) a. Determines the physical configuration and infrastructure of the supply chain. The supply chain infrastructure needs to be reevaluated due to changes in demand patterns, product mix, production process etc. The following key strategic decisions are integral: i. Determining the appropriate no. of facilities ii. Determining location of facility iii. Determining size of facility iv. Allocating space for pdts per facility v. Determine sourcing requirements vi. Determine distribution strategies b. Objective = MINIMIZE systemwide costs!!!!! c. Tradeoffs of :

ii. Increase in inventory costs due to iii. Increase in overhead and setup costs iv. Reduction in outbound tpt costs, but increase in inbound transportation costs. v.  in essence : must balance the costs of opening new warehouses with the advantages of being close to customers. d. Data collection: i. Problem involves large amounts of data on: 1. Locations of customers, retailers etc 2. All products 3. Annual demand 4. Transportation rates by mode 5. Warehousing costs (labour, inventory carrying charges etc)

6. 7. 8. 9.

Shipment sizes Order processing costs Customer service requirements/goals Production and sourcing costs/capabilities.

ii. Procedure: First, customers located in close proximity to each other are aggregated using a grid network/other clustering technique (customer zones) iii. Next, Items are aggregated into a reasonable number of product groups, based on: 1. Distribution pattern 2. Product type iv. Important consideration – the impact on the model’s effectiveness of replacing the original detailed data with the aggregated data. Addressed in 2 ways: 1. Even if the technology exists to solve the logistics network design problem with the original data, may still be usefl to aggregate data because our ability to forecast is usuall poor 2. Aggregating customers into 150-200 zones usually results in no more than a 1% error of total transportation costs. In practice, the following approach is typically used: -

make sure each zone has approximately an equal amount of total demand place the aggregated points at the centre of the zone a –

STEP 2 of NETWORK DESIGN: ESTIMATE TRANSPORTATION COSTS -

-

-

an important characteristic of transportation rates, is that rates are almost linear with DISTANCE but not with volume! Distinguish between transportation costs associated with and internal and external fleet. Also, distinguish between two modes of transport: truckload (TL) vs less than truckload (LTL). o An important property of TL costs is that it is not symmetric i.e. it is typically more expensive to ship a fully loaded truck from A to be, than it is from B to A. TL – TL carriers subdivide the country into zones, and provide their clients with zone-to-zone table costs. LTL – rates typically belong to one of the 3 basic types of freight rates: class, exception and commodity. o In all cases, the higher the rating or class, the greater the relative charge for transporting the commodity. o once the rating is establish, it is important to identify the rate basis number. (the approximate distance between the load’s origin and destination)

-

the other 2 freight rates – exception and commodity are specialized rates used to provide either less expensive rates (exception) or commodity-specific rates (commodity).

Mileage Estimation -

The straight line distance in miles from a to b: 𝐷𝑎𝑏 = 69√(𝑙𝑜𝑛𝑎 − 𝑙𝑜𝑛𝑏 )2 − (𝑙𝑎𝑡𝑎 − 𝑙𝑎𝑡𝑏 )2 𝑤ℎ𝑒𝑟𝑒 69 = 𝑛𝑜. 𝑜𝑓 𝑚𝑖𝑙𝑒𝑠 𝑝𝑒𝑟 𝑑𝑒𝑔𝑟𝑒𝑒 𝑜𝑓 𝑙𝑎𝑡𝑖𝑡𝑢𝑑𝑒 to measure with respect to curvature: INSER EQUATION!

To correct for underestimations of road distance (as both factors underestimate the actual road distance, multiply by CIRCUITY FACTOR!) Warehouse costs -

-

typically consists of 3 components: o handling costs – labour and utility costs o fixed costs – all components that are not proportional to amount of material that flows through the warehouse. o storage costs – inventory holding costs. Estimating the warehouse handling costs is fairly easy but the other 2 are quite difficult. – annual flow and average inventory associated with the product tells us nothing about how much space is need for the product. Hence, to actually overcome this, use inventory turnover ratio!

Warehouse capacities  inventory turnover is a useful way to have an appropriate approach. Annual flow is divided by the inventory turnover ratio to calculate the average inventory level. Assuming a regular shipment and delivery schedule, it follows that the required storage space is approximately TWICE that amount (as avg = Q/2). We then multiply this by a factor, of 3. i.e. where annual flow = 1000, inventory turnover ratio is 10, this implies that avg inventory 100, and if each unit takes 10 sq feet, total space required for warehouse = 100 * 2 * 10 *3 = 6000 (where required space for pdts = 2000) Potential warehouse locations -

important to effectively identigy potential locations for new warehouses. Must satisfy some conditions: Geographical and infrastructure conditions, natural resources and labour availability, local industry and tax regulations, public interest.

Model and data validation -

-

typically done by reconstructing the existing network configuration using the model and data collection, and comparing the output of the model to existing data. The importance of validation cannot be overstated!

-

Also useful to use a what-if analysis, including: o Does the model make sense? o Are the data consistent? o Can the model results be explained? o Did u perform sensitivity analysis

Solution techniques: once the data are collected, tabulated and verified, next step is to optimize the configuration of the logistics network: 2 techniques 1) Mathermatical optimization techniques (heuristics etc) a. These tools can determine strategies that will significantly reduce the total system costs! b. However, they are static models that do not take into account changes over time. 2) Simulation models that provide a mechanism to evaluate specified design alternatives. a. However, the onkly model a prespecified logistics network. i.e. given a particular config of warehouses, retailers etc, a simulation model can be used to help estimate the costs associated with operating that configuration. b. Typically, one can consider very few alternatives using a simulation tool. 3) Typically, use a mathematical optimization model the generate a number of least-cost solutions at the macro level, and use a simulation model to evaluate the solutions generated in the first phase! (always take note of not only the EFFECTIVENESS, but also the ROBUSTNESS of the supply chain planning tool!) Inventory positioning and Logistics Coordination Strategic Safety Stock -

-

One ...