Task 879.2: Integration of Demand Planning and Manufacturing - PowerPoint PPT Presentation

SRC Project 879 Progress report Task 879.2: Integration of Demand Planning and Manufacturing Planning Task leader: Yon-Chun Chou Co-PI: Argon Chen National Taiwan University 2001.10.29

Y-C Chou Product Variety Granularity (product hierarchy level) Specific Product products Variety Generic 1, products L1 2, L2 3, α β γ , , L3 ... Time (month) 3 6 9 H1 H2 H3 (horizon)

Y-C Chou Multi-granular Capacity Planning p uncertainty d : demand forecast made at time p i , t Tool for product i in time t require. ⇓ derive ( k ) 1 R k , 1 1 p R R : capacity requiremen ts of tool k k , 2 1 R k , t k , 4 2 R k , 1 2 2 R R k , 2 k , 3 Time ( t ) 1 2 Based on level 2 information Planning Based on level 1 information H1 time ( p ) H2 They are estimates for capacity requirement of the same time period.

Y-C Chou Research Needs and Objectives • Needs: To develop capacity plans based on demand forecasts for aggregate products (such as α ). (It may be too late to wait for level 1 demand information.) Objectives: • Develop a quantitative model for the uncertainty in capacity requirements • Identify the tools that are sensitive to changes in demand scenarios

Y-C Chou The Envelope of Capacity Requirements → d ( d , d ) α 1 2 = ⋅  d a d α 1 1 ≤ ≤ + = 0 a , a 1 and a a 1  1 2 1 2 The envelope of = ⋅ d a d  α 2 2 capacity requirements demand scenarios : ( a , a ) 1 2 of all scenarios Tool 2 E1 = = E : a 1 , a 0 1 1 2 = = E : a 0 , a 1 2 1 2 E2 Tool 1 This envelope has been proven mathematically.

Y-C Chou A Numerical Example • Total demand d α = 20,000 wpm, d α = d 7 + d 8 Product Products Group α 7, 8 Demand Scenario 1 2 3 4 5 6 7 8 9 10 11 Product 7 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Product 8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 E1 E2 * Process flow data is provided by TSMC.

Y-C Chou Tool Requirements of All Tool Sets • Tools 15, 18, 33, 46, and 54 are sensitive to demand scenario. Scenario1 14 Scenario2 12 Tool requirement Scenario3 10 Scenario4 8 Scenario5 6 Scenario6 4 Scenario7 2 Scenario8 Scenario9 0 Scenario10 0 20 40 60 80 100 Tool ID Scenario11

Y-C Chou Tool Requirements of Sensitive Tool Sets • Scenarios 1 and 11 are the extremes, their tool requirements enclose those of all other scenarios. 16 14 Scenario1 Tool requirement 12 Scenario2 Scenario3 10 Scenario4 8 Scenario5 Scenario6 6 Scenario7 4 Scenario8 2 Scenario9 Scenario10 0 Scenario11 0 10 20 30 40 50 60 Tool ID

Y-C Chou Validating the Envelope of Capacity Requirements scenarios E1 7 6 5 Tool 33 4 3 2 1 E2 0 0 5 10 15 Tool 46

Y-C Chou The Number of Extreme Demand Scenarios = 3 The envelope is a region. → d ( d , d , d ) α 1 2 3 = ⋅  d a d α 1 1  = ⋅ d a d  α 2 2 Tool 2  E3 = ⋅ d a d  α 3 3 E1 ≤ ≤ + + = 0 a , a , a 1 and a a a 1 1 2 3 1 2 3 scenarios : ( a , a , a ) 1 2 3 E2 Tool 1

Y-C Chou Numerical Example: 3 Extreme Scenarios Product Products Group α 7, 8, 9 Scenario 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Product 7 1.00 0.75 0.75 0.50 0.50 0.50 0.25 0.25 0.25 0.25 0.00 0.00 0.00 0.00 0.00 Product 8 0.00 0.25 0.00 0.50 0.00 0.25 0.75 0.00 0.50 0.25 1.00 0.00 0.50 0.25 0.75 Product 9 0.00 0.00 0.25 0.00 0.50 0.25 0.00 0.75 0.25 0.50 0.00 1.00 0.50 0.75 0.25 E1 E2 E3

Y-C Chou Validating the Capacity Envelope: 3 Extremes 12 10 8 Tool 46 6 4 2 0 0 5 10 15 Tool 33

Y-C Chou Two Product Groups, 5 Products Product Products Group α 7, 8, 9 β 25, 28 12 10 8 Tool 46 6 4 2 0 0 5 10 15 Tool 33

Y-C Chou The Number of Tool Sets = 3 Tool 2 The envelope is a convex hull. Tool 1 Tool 3 Assuming 4 extreme demand scenarios

Y-C Chou A Confidence Model of Capacity Insufficient Tool 2 Tool requirements covered Total area = A by all scenarios A A 3 4 Tool 2 A A 1 2 n 2 + + A 0 . 5 ( A A ) 1 2 4 = confidence A n 1 Tool 1 Provisioned Capacity = tool portfolio (n 1 , n 2 )

Y-C Chou Work in Progress • Constructing the convex hull (in n-D space) • Computing the “volume” of convex hull algorithms • Decomposing the convex hull Will be able answer these questions: • If the quantity of tool set x is increased by 1, what is the improvement in “confidence”? • Which tool set is the most sensitive to demand scenario? → → d tool requiremen ts R capacity α α → product hierarchy and demand scenario confidence

Y-C Chou Questions • In the above analysis, demand scenarios are continuous. An alternative approach is to model demand scenarios as discrete. • Improving the confidence model?