Inventory Management Extensions to EOQ Chris Caplice ESD.260/15.770/1.260 Logistics Systems Oct 2006 Agenda Review of Basic EOQ Non-instantaneous Leadtime Finite Replenishment (EPQ) Multiple Locations Discounting MIT Center for Transportation & Logistics – ESD.260 2 © Chris Caplice, MIT Economic Order Quantity (EOQ) Finding the order quantity Q (and frequency T) that minimizes the total relevant cost. AD vrQ TRC[Q] = + Q 2 2 AD Q* = vr T* = 2A Dvr TRC* = 2 ADvr MIT Center for Transportation & Logistics – ESD.260 3 © Chris Caplice, MIT Assumptions: Basic EOQ Model Demand Discounts Constant vs Variable Known vs Random Continuous vs Discrete Instantaneous Constant or Variable (deterministic/stochastic) Dependence of items Independent Correlated Indentured Continuous Periodic One Multi (>1) None Uniform with time Single Period Finite Period Infinite One Many Single Stage Multi-Stage Form of Product Unlimited Limited / Constrained MIT Center for Transportation & Logistics – ESD.260 Number of Items Capacity / Resources None All orders are backordered Lost orders Substitution Planning Horizon Number of Echelons Perishability Review Time None All Units or Incremental Excess Demand (Shortages) Lead time 4 © Chris Caplice, MIT Extensions: Leadtime Order Leadtime Demand Positive (nonzero) Deterministic Impact Order Receive Does this change Q*? What is my new policy? What is my new avg IOH? L = Order Leadtime Inventory On Order Inventory On Hand MIT Center for Transportation & Logistics – ESD.260 EOQ Inventory Policy: Order Q* units when IOH = DL 5 © Chris Caplice, MIT Extensions: Finite Replenishment Inventory becomes available at a rate of m units/time rather than all at one time Does this change Q*? What is my new policy? What is my new avg IOH? Q(1-D/m) Inventory On Hand Slope = m-D ⎛ D⎞ Q ⎜1 − ⎟ vr AD m⎠ + ⎝ TRC[Q] = 2 Q Slope = -D MIT Center for Transportation & Logistics – ESD.260 ⎛ ⎛ D⎞⎞ 2 AD EPQ = = EOQ ⎜ ⎜1 − ⎟ ⎟ ⎜ ⎝ m⎠⎟ ⎛ D⎞ ⎝ ⎠ vr ⎜1 − ⎟ ⎝ m⎠ 6 −1 © Chris Caplice, MIT Extensions: Multiple Locations Suppose that instead of one location satisfying all demand, there are n locations. Each location serves di = D/n units of demand Identical (uniform) demand at each location What is my new inventory policy? What is my new average Inventory on Hand? How much is this better or worse than a single location? Questions Q* 1 1 2 AD Q* = vr Q* IOH = TRC* = 2 DAvr 2 MIT Center for Transportation & Logistics – ESD.260 q1* 2 3 qi* = versus 2 Adi 2 AD = vr vrn n ⎛q IOH = ∑ i =1 ⎜ ⎝ 2 * i 7 q2* q3* TRC* = 2nDAvr ⎞ ⎛ Q*⎞ = n ⎟ ⎜ ⎟ ⎝ 2 ⎠ ⎠ © Chris Caplice, MIT Extensions: Multiple Locations What if I reduce number of stocking locations from M to N? M TRC*[M ] 2MDAvr = = N TRC*[ N ] 2 NDAvr What if my sub-regions do not have uniform demand? Is this a reduction in safety stock, cycle stock, or both? How dependent is this effect on inventory policy at each site? EOQ Policy (order qEOQ* when IOHi=0) Fixed Order Size (Always order a full truckload at a time) Days of Supply (Always order a month’s supply) MIT Center for Transportation & Logistics – ESD.260 8 © Chris Caplice, MIT Extensions: Multiple Locations Fixed Order Size, e.g. only order full truckloads Days of Supply, e.g. order at start of each month For a Single Location Policy EOQ FOS DOS Q* QFOS QDOS Order Size Average IOH Q*/2 QFOS/2 QDOS/2 Order Cost OEOQ OFOS ODOS Holding Cost HEOQ HFOS HDOS OEOQ + HEOQ OFOS + HFOS ODOS + HDOS Total Cost For N Locations Policy EOQ FOS DOS q* QFOS qDOS Average IOH √N(Q*/2) N(QFOS/2) QDOS/2 Order Cost √N( OEOQ) OFOS N(ODOS) Holding Cost √N (HEOQ) N(HFOS) HDOS √N (OEOQ+HEOQ) OFOS+NHFOS NODOS+HDOS Order Size Total Cost MIT Center for Transportation & Logistics – ESD.260 9 Example DATA A= 500 D = 2000 r= 0.25 v= 50 N= 4 Trk Cap = 500 DOS = 0.083 30 $/order Units/year $/$/year $/unit locations units/shipment years days Single Location Policy Order Size Average IOH Order Cost Holding Cost Total Cost EOQ FOS 400 500 200 250 $ 2,500 $ 2,000 $ $ 2,500 $ 3,125 $ $ 5,000 $ 5,125 $ DOS 4 Locations Policy Order Size Average IOH Order Cost Holding Cost Total Cost EOQ FOS 200 500 400 1000 $ 5,000 $ 2,000 $ $ 5,000 $ 12,500 $ $ 10,000 $ 14,500 $ DOS 167 83 6,000 1,042 7,042 42 21 24,000 1,042 25,042 © Chris Caplice, MIT Extensions: Discounts All Units Discount Discount applies to all units purchased if total amount exceeds the break point quantity Examples? Incremental Discount Discount applies only to the quantity purchased that exceeds the break point quantity Examples? One Time Only Discount Less common – but not unheard of! A one time only discount applies to all units you order right now (no quantity minimum or limit) How will different discounting strategies impact your lot sizing decision? What cost elements are relevant? MIT Center for Transportation & Logistics – ESD.260 10 © Chris Caplice, MIT Extensions: All Units Discounts Two Cases to Examine . . . ⎧v0 v=⎨ ⎩ v0 (1 − d ) 0 ≤ Q ≤ Qb Qb ≤ Q AD v0 rQ ⎧ Dv + + 0 ≤ Q ≤ Qb 0 ⎪ Q 2 ⎪ TRC = ⎨ ⎪ Dv (1 − d ) + AD + v0 (1 − d ) rQ Qb ≤ Q ⎪⎩ 0 Q 2 Qb $112,000 $110,000 Where d = Discount v0 = Base unit price Qb = Break quantity Total Cost $108,000 $106,000 $104,000 $102,000 Typically, Q* < Qb but what if Q* > Qb? $100,000 $98,000 100 200 300 400 500 600 700 800 900 1000 Order Quantity (Lot Size) TC2 MIT Center for Transportation & Logistics – ESD.260 TC1 11 © Chris Caplice, MIT Extensions: All Units Discounts Simple efficient algorithm 1. 2. 3. Find EOQ with discount (EOQd) If EOQd≥Qb then pick EOQd Otherwise, go to 3 Solve for TRC(Q*) and TRC(Qb) If TRC(Q*) < TRC(Qb) then pick Q* Otherwise, pick Qb Can be extended to more than one break point Example: D=2000 Units/yr r=.25 A=$500 v0 = $50 Discount of 2% off if Q≥500 MIT Center for Transportation & Logistics – ESD.260 12 © Chris Caplice, MIT Extensions: Incremental Discounts Total Purchase Cost Discount only applies to quantity above breakpoint Trade-off between lower purchase cost and higher carrying costs Cost of units ordered below each breakpoint are essentially ‘fixed’ v1 v2 v0 v1Q1 v0Q1 (v0-v1)Q1 Q1 MIT Center for Transportation & Logistics – ESD.260 Q2 13 Units © Chris Caplice, MIT Extensions: Incremental Discounts Efficient algorithm 1. 2. 3. 4. Find Fixed Cost per breakpoint, Fi, for each break Find EOQi for each range – including the Fi If EOQi is not within allowable range, go to next I Otherwise, find TRCi using effective cost per unit, vei Pick EOQi with lowest TRC Can be extended to more than one break point Fi = Fi −1 + ( vi −1 − vi ) Qi F0 = 0 2 D ( A + Fi ) EOQi = rvi MIT Center for Transportation & Logistics – ESD.260 vi EOQi + Fi v = EOQi e i 14 © Chris Caplice, MIT Example: Incremental Discounts Price Breaks: 10% off for 500 to <1000 units 20% off for 1000 or more units i=2 i=1 i=0 vi $40 $45 $50 Qbi 1,000 500 0 Fi 7,500 2,500 0 EOQi 1,789 OK 1,033 X 400 OK vei $44.19 Purch Order Hold TRCi D=2000 Units/yr r=.25 A=$500 v0 = $50 $50 $ 88,384 $ 559 $ 9,882 $ 98,825 MIT Center for Transportation & Logistics – ESD.260 $100,000 $ 2,500 $ 2,500 $ 105,000 15 © Chris Caplice, MIT Extensions: One Time Discount Suppose you are offered a One Time deal! Should you take it? vg = One time good deal purchase price ($/unit) Qg = One time good deal order quantity (units) TCsp=TC over time covered by special purchase ($) MIT Center for Transportation & Logistics – ESD.260 16 © Chris Caplice, MIT Extensions: One Time Discount Compare Options: Not Special Price vs. Special Price Find TC for normal price TC = (CycleTime)(TC * + PurchaseCost ) ⎛ Qg ⎞ ⎛ Qg TC = ⎜ ⎟ 2 ArvD + ⎜ ⎝ D⎠ ⎝ D ⎞ ⎟ vD ⎠ Find the Savings (TC-TCSP) Savings = TC − TCSP ⎛ ⎛ Qg ⎞ ⎛ Qg = ⎜⎜ ⎟ 2 ArvD + ⎜ ⎝ D ⎝⎝ D ⎠ MIT Center for Transportation & Logistics – ESD.260 ⎞ ⎞ ⎞ ⎛ ⎛ Qg ⎞⎛ Qg ⎞ ⎟ vD ⎟ − ⎜ vg Qg + rvg ⎜ ⎟⎜ ⎟ + A⎟ ⎠ ⎠ ⎝ ⎝ 2 ⎠⎝ D ⎠ ⎠ 17 © Chris Caplice, MIT Extensions: One Time Discount Finding 1st and 2nd order conditions (Maximize Savings) ⎛ 2rvg Qg ⎞ dS ⎛1⎞ = 0 = ⎜ ⎟ 2 AvrD + ( v − vg ) − ⎜ ⎟ d (Qg ) D 2 D ⎝ ⎠ ⎝ ⎠ ⎛ 2rvg d 2S = −⎜ 2 d (Qg ) ⎝ 2D So that the Optimal Quantity to buy is ⎛ D Q =⎜ ⎜ Drv g ⎝ * g ⎞ ⎟<0 ⎠ Cleaning this up gives: MIT Center for Transportation & Logistics – ESD.260 D ( v − vg ) ⎞ ⎟⎟ 2 ArvD + rvg ⎠ ⎛ v Q =Q ⎜ ⎜v ⎝ g * g 18 * ⎞ D ( v − vg ) ⎟⎟ + rvg ⎠ © Chris Caplice, MIT Take-Aways EOQ is a good place to start for most analysis EOQ can be extended to cover Non-zero leadtimes Finite replenishment systems Multiple locations Square Root law rests on implicit assumptions Distribution of demand and inventory policy will impact results Discounts Purchase price (v) becomes relevant Common in practice (economies of scale) MIT Center for Transportation & Logistics – ESD.260 19 © Chris Caplice, MIT Questions? Comments? Suggestions?