# ISEISE-402

Production Systems
Lecture on
Aggregate Planning

Chapter 3 Aggregate Planning
Presented By: Professor Abdulbasit A. Andijani Systems Engineering Department
SE 402 Production Systems

• Micro planning and macro planning • Use to determine
– number of employees (technicians/nurses/flight attendance/pilots/delivery drivers/teachers) – quantity and product mix • Main assumption: demand is deterministic p

• The objectives
– Meet changing demand (changes in the size of g g ( g manpower) – Maintain a stable work force ( skilled labor ) – Maximize profit

How to aggregate ? Unit of production Weight (ton of cement) g g Volume ( gallons of engine oil ) Man-hours \$ of inventory

Example
Model A B C D E Man-hrs 3 3.5 4 4.5 45 6 % of sales 45 25 15 10 5 Selling \$ 210 280 400 540 900

• • • •

Same product family Can t C not use selling price (why ?) lli i ( h Can use man-hrs ( average man-hrs ) g Aggregate production unit:

an imaginary product requiring the average man hours • A Aggregate production unit = t d ti it ∑ forecast of each type*its % of sales

If there is a large variety of g y products. selling price is used and it gives a good approximation .

Input: Demand forecast for product or service per unit time .

We ill W will assume no forecasting error!! f ti !! .We must have forecast of aggregate demand for the planning horizon.

Determine level of resources (manpower) Objective: maximize profit .Output: 1. Determine aggregate production quantity in each planning period i i h l i i d 2.

No need for AP if demand is stable and resources are infinite! .

planning horizon 4. bottleneck problems 2 b ttl k bl 3.Main issues in aggregate planning: 1. smoothing 2. treatment of demand .

1. 1 Smoothness: Refers to cost of hiring and firing and drawbacks of each .

2 Bottleneck problems: Refers to large increases in the demand that can not be handled .2.

3 Planning horizon: * Short horizons result in not meeting demand later * Long horizons result in inaccurate forecasts * Effect of end-of-horizon inventory * Use of rolling horizons * Long set up times result in frozen time periods .3.

Treatment of demand 4 T t t fd d No forecasting errors N f ti .4.

5. Hiring costs g y Firing/lay off cost Holding cost g Shortage cost Regular time cost g Overtime/ Subcontracting cost Idle time cost . 6. 4.Costs involved 1. 3. 7. 2.

1. Hiring costs Time of recruiting team i f ii Ads. Firing/lay ff 2 Fi i /l off cost Severance pay Poor worker morale P k l Will assume a linear function . interviews Training T i i Will assume a linear function 2.

backlogged demand (common in manufacturing) g) b. lost demand ( common in retail) Will assume a linear function . Holding cost Cost due to having inventory It is based on a. level of inventory during the planning period or end of period Will assume a linear function 4. unit value of inventory it l fi t b. Shortage cost Cost due to shortage a.3.

Overtime and subcontracting cost Pay to regular workers at overtime Cost paid to outside suppliers We assume a linear function . Regular time cost Manpower Manpo er Direct and indirect material cost We assume a linear function g 6.5.

7. Idle time cost Usually U ll zero .

Example Month Demand Notes 1 1280 500 available 2 640 3 900 4 1200 5 2000 6 1400 600 ending invt .

Example Month Demand 1 1280 2 640 3 900 4 1200 5 2000 6 1400 Net Demand 780 640 900 1200 2000 2000 .

How to include: minimum buffer of 200 i each month i i b ff f in h th Add the minimum buffer to the demand of the first month .

Example Month Demand 1 1280 2 640 3 900 4 1200 5 2000 6 1400 Net Demand 980 640 900 1200 2000 2000 .

How to include a minimum bu e o 00 buffer of 200 in month 3? o t Add the minimum buffer to the demand of month 3 and subtract from month 4 .

Example Month Demand 1 1280 2 640 3 900 4 1200 5 2000 6 1400 Net Demand 780 640 1100 1000 2000 2000 .

Example Cost f hiring C of hi i SR 500/ 500/worker k Cost of firing SR 1000/worker Cost f holding C of h ldi 1 unit of inventory i fi for 1 month SR 80 .

14653 .76 workers produced 245 units in it i 22 days Worker productivity/day = 245/( 76 * 22) = 0 14653 0.

Zero inventory plan Chase the demand plan Initial workforce is 300 Demo .

4 78.765 worker productivity Initial work force Firing cost Hiring cost g Storage cost 0.420 2.80981 3.51675 2.260 .000 2.760 Total cost of plan 524.1465 0 1465 300 1000 500 80 2.93062 2 93062 3.000 Production 782 640 902 1.63756 3.320 3.5 511.000 Extra Quantity 2 0 2 0 1 0 5 Cumulative Production 782 1.19797 Forecasted Month Jan Feb March April May June demand 780 640 900 1.200 2.520 Total Cost Inventory 2 2 4 4 5 5 22 1.000 Working days 20 24 18 26 22 15 No of workers 267 182 342 315 621 910 Total Cost Workers Hired 0 0 160 0 306 289 755 377.500 Workers Fired 33 85 0 27 0 0 145 145.525 7.324 3.525 Cumlative Demand 780 1.22368 2.001 2.Zero inventory plan p Chase the demand plan Chase demand or zero inventory plan sensitivity analysis 1050.524 5.520 7.200 2.422 2.520 5.

Note the difference between Inventory and Extra quantity y q y 3. Labor cost is not included 2. Frequent hiring and firing 4.Remarks R k 1. 4 Include the holding cost of the ending inventory .

The case of constant work force ork How large is it? Demo .

146531 p y Initial work force 300 Firing cost 1000 Hiring Hi i cost t 500 Storage cost 80 How many to hire ? No shortage is allowed Workers needed Forecasted Working Cumulative Cumulative Month demand days demand work days Jan 780 20 780 Feb 640 24 1420 March 900 18 2320 April 1200 26 3520 May 2000 22 5520 June 2000 15 7520 y The constant workforce which will satisfy 20 267 44 221 62 256 88 273 110 343 125 411 the demand is 411 .worker productivity 0.

worker productivity 0.1465 Initial work force Firing cost Hiring cost Storage cost 300 1000 500 80 Forecasted Working No of Cumulative Cumulative Month demand days workers Production Production Demand Inventory Jan Feb March April May June 780 640 900 1200 2000 2000 Labor cost 55500 20 24 18 26 22 15 411 411 411 411 411 411 1204 1445 1084 1565 1324 903 1204 2649 3733 5298 6622 7525 780 1420 2320 3520 5520 7520 Inventory Cost 424 1229 1413 1778 1102 5 5951 476080 Total cost of plan 531580 .

8000 7000 6000 5000 4000 3000 2000 1000 0 1 2 3 4 5 6 .

Labor cost is not included 2.Remarks 1. Hiring or firing occurs only once 4. 2 Excessive inventory cost 3. Include the holding cost of g y the ending inventory .

4 Guarantees optimal solution 5. Cost parameters may be time varying values . More constraints could be added 3. For larger problems 2.Linear Programming 1. Regular and overtime labor cost can easily g y be included 4.

Decision variables Let T be the planning horizon Wt = workforce in period t Ht = workers hired in period t Ft = workers fi d in period t k fired i i d Pt = units produced during period t (during regular and overtime) Ot = units produced during overtime Ut = idle time in period t measured in product units St = units supplied by a subcontractor It = i inventory level in period t l li i d .

Parameters Dt = demand in period t CH = cost of hiring one worker CF = cost of firing one worker CI = cost of holding 1 unit for 1 period CR = cost of producing 1 unit in regular time CO= cost of producing 1 unit in overtime CS = cost of 1 unit supplied by the subcontractor f i li d b h b nt = number of working days in period t K = number of units a worker can produce in a day W0 = initial workforce .

The constraints Conservation of workers constraints No of No of Number workers workers hired in = + in period in period period t t–1 t Wt = Wt –1 + Ht Number fired in period t Ft .

Conservation of units constraints Inventory in period t = = Inventory in period t–1 + + Units i produced in period t + + Units subcontracted in period t - Demand in period t Dt It It –1 Pt St .

Production constraints Units Idle time Production produced Units capacity in units in produced = in regular + of in period overtime time in i i production t time in t t t Pt = KntWt + Ot Ut .

T [ H Ht + cF Ft + cI It [c + cR (Pt – Ot) + cO Ot + cS St + cU Ut ] .Objective function Minimize ∑t=1.

Solution may not be integer Can we get Ot > 0 and Ut > 0 Ht > 0 and Ft > 0 ? .

Solution f h S l i of the problem using linear bl i li programming .

6375598094 W3 = 0 P4 .W1 .2.3.I1 + P1 = 780 I1 .W5 .I4 + P4 = 1200 I4 .I2 + P2 = 640 I2 .223684211 W5 = 0 P6 .2.3.H2 + F2 = 0 W3 .W3 .I5 + P5 = 2000 I5 – I6 + P6 = 1400 P1 .2.516746411 W2 = 0 P3 .809808612 W4 = 0 P5 .93062201 W1 = 0 P2 .H6 + F6 = 0 I6 = 600 .3.H1 + F1 = 300 W2 .197966507 W6 = 0 END GIN W1 GIN W2 GIN W3 GIN W4 GIN W5 GIN W6 .W2 .W4 .H4 + F4 = 0 W5 .I3 + P3 = 900 I3 .H3 + F3 = 0 W4 .MIN 500 H1 + 1000 F1 + 80 I1 + 500 H2 + 1000 F2 + 80 I2 + 500 H3 + 1000 F3 + 80 I3 + 500 H4 + 1000 F4 + 80 I4 + 500 H5 + 1000 F5 + 80 I5 + 500 H6 + 1000 F6 + 80 I6 ST W1 .H5 + F5 = 0 W6 .

This problem has no feasible solution? Why? .

I3 + P3 I3 .I5 + P5 I5 – I6 + P6 = 780 = 640 = 900 = 1200 = 2000 = 1400 -600 + ∑ Pt = 7520 → ∑ Pt is integer .I1 + P1 I1 .I2 + P2 I2 .The Th sum of the inventory constraints: f h i i I6 = 600 .I4 + P4 I4 .

516746411 W2 = 0 P3 .3.93062201 W1 = 0 P2 .809808612 W4 = 0 P5 .223684211 W5 = 0 P6 .3.From th manpower constraints: F the t i t P1 .2.2.6375598094 W3 = 0 P4 .3.197966507 W6 = 0 ∑ Pt is fractional .2.

We add the conditions: Pt ≤ kntWt and Pt + 1 ≥ kntWt .

809808612 W4 <= 1 + 3.2.3.197966507 W1 <= 0 W2 <= 0 W3 <= 0 W4 <= 0 W5 <= 0 W6 <= 0 + 2.P1 P2 P3 P4 P5 P6 -P1 -P2 -P3 -P4 -P5 -P6 .6375598094 W3 <= 1 + 3.3.809808612 .93062201 W1 <= 1 + 3.2.516746411 W2 <= 1 + 2.223684211 .6375598094 .197966507 W6 <= 1 .223684211 W5 <= 1 + 2.3.2.516746411 .93062201 .

901 600 Total cost = 378.05 273 0 0 959.897.07 338.Period 1 2 3 4 5 6 Labor Hired Fired Production Inv 273 0 27 799.24 737 462 0 2375.05 4.5 .855 380.1 737 0 0 1619.19 157 19 275 2 0 1047.05 719 05 157.05 19.13 273 0 0 719.

Extensions: Minimum buffer Limits of hiring/firing Production capacity Backlog of excess demand .

097 .What if: Maximum production ≤ 2000 Period Labor Hired Fired P i d L b Hi d Fi d 1 2 3 4 5 6 300 300 300 377 620 620 0 0 0 77 243 0 0 0 0 0 0 0 Cost = 391.

111 .2 Wt-1 Period Labor Hired Fired P i d L b Hi d Fi d 1 2 3 4 5 6 300 309 370 444 532 534 0 9 61 74 88 2 0 0 0 0 0 0 Cost = 420. Wt ≤ 1.In addition.948. what if: Maximum manpower increase is 20% of previous value. I4 = 1.

1 .In addition.884. what if: inventory can not exceed 1000. It ≤ 1000 Period Labor Hired Fired P i d L b Hi d Fi d 1 2 3 4 5 6 300 300 359 430 516 610 0 0 59 71 86 94 0 0 0 0 0 0 Cost = 428.

7 The linear decision rule g g 3.Skip 3.9 Disaggregating aggregate plans .8 Modeling management behavior 3.