LOGI0011-1 : Supply Chain Management Group Project - Group 2 2020-2021 Group 2G LIMBOURG Sabine - BEBRONNE Élodie CANTINEAU Simon HERMESSE Charles YAMADJAKO Mathieu Exercise 1 It is a capacited plant allocation model. i ∈ {1 = New York, 2 = Savannah, 3 = Cleveland, 4 = San Diego} j ∈ {1 = East, 2 = South, 3 = Midwest, 4 = West} Parameters - dj = Demand from market j, ∀j. - ki = Potential capacity of factory i, ∀i. - fi = Fixed costs of keeping factory in i open, ∀i. - cij = Cost of producing and shipping a sunshade from factory i to market j, ∀i and ∀j. Variables - xij = Quantity shipped from factory i to market j, ∀i, j. 1 if a high − capacited factory π is open, - yHi = { ∀i 0 if a high − capacited factory π is not open 1 if a low − capacited factory π is open, - yLi = { ∀i 0 if a low − capacited factory π is not open Objective function 4 4 4 min ∑ ∑ πππ π₯ππ + ∑ ππ π¦π π=1 π=1 π=1 Constraints 4 π·πππππ πΆπππ π‘πππππ‘: ∑ π₯ππ ≥ ππ ∀π π=1 π΅ππππππ‘π¦ πΆπππ π‘πππππ‘: π¦π ∈ {0,1} ∀π 4 ππ‘πππππ πΆππππππ‘π¦ πΆπππ π‘πππππ‘: ∑ π₯ππ ≤ 210 000 × π¦πΏπ + 390 000 × π¦π»π ∀π π=1 πππ − πππππ‘ππ£ππ‘π¦ πΆπππ π‘πππππ‘: π₯ππ ∈ β€+ ∀π, π The optimal solution costs $131,403,000. We open a small plant in New York and a large plant in Cleveland. We don’t open any plant in San Diego and in Savannah. 1 Exercise 2 2.1. Parameters - dt = Demand forecast for month t, ∀t ∈ {1, …, 18} - Tt = Number of working days in month t, ∀t ∈ {1, …, 18} - I0 = 900 = Aggregate inventory at the beginning of month 1. - W0 = 23 = Number of workers in June Variables - Xt = Quantity of unit produced in month t - It = Quantity in stock at the end of month t. - St = Quantity of unit subcontracted in month t. - Wt = Number of employees for month t. - Ht = Number of employees hired at the beginning of month t. - Lt = Number of employees laid off at the beginning of month t. Objective function 18 18 18 min (1250 × ∑ π»π‘ + 1400 × ∑ πΏπ‘ + 10 × 8 × ∑ ππ‘ × ππ‘ + 3,2 18 π‘=1 18 π‘=1 18 π‘=1 × ∑ πΌπ‘ + 0,4 × ∑ ππ‘ + 35 × ∑ ππ‘ ) π‘=1 Constraints Inventory balance constraint: Inventory balance constraint: Workforce balance constraint: Capacity constraint: Xt, It, St, Wt, Ht, Lt ∈ β€+ π‘=1 π‘=1 I18 = 1800 It = It-1 + Xt + St – dt ∀t ∈ {1, …, 18} Wt = Wt-1 + Ht – Lt ∀t ∈ {1, …, 18} π ×π Xt ο£ π‘0,3 π‘ ∀t ∈ {1, …, 18} ∀t ∈ {1, …, 18} The minimized cost to plan the production from July to March is $1,554,458.4 using a tolerance of 1%. 2.2. By using the solver and trying a different subcontracting price, we noticed that below $23.1, it was worth subcontracting because it was cheaper than producing units ourselves. Indeed, as we can see in the table, when the subcontracting price is below $23.1, the number of subcontracted units remains the same. Otherwise, when the price is higher than $50.0736, we have to stop subcontracting because it is more expensive than our production costs. Thanks to the solver, we can see that when the price is higher than $50.0736, the subcontracted units are zero. 2 Evolution of the number of subcontracted units according to price 65000 60000 Number of subcontracted units 55000 50000 45000 40000 35000 30000 25000 20000 15000 10000 5000 0 -5000 20 25 30 35 40 Price of a subcontracted unit 45 50 3 Exercise 3 ο = 285 units/day ο³ = 28 units/day A = $9 h = $0.35 units/year 0.35 ο h = 365 = $9.589x10-4 units/day L = 7 days 3.1 2×π΄×π Q* = √ β = 2,312.97 ο» 2,313 units R* = ο*L = 1995 units The economic order quantity for the bar of soap is 2,313, meaning that when the company places an order, it orders 2,313 bars of soap. The optimal reorder point is 1,995 units. Thus, when the stock drops below 1,995 bars of soap, the company needs to place a new order. 3.2 ο‘ = 99% ο Pr(z ο£ k) = 0,99 ο k = 2.33 π −π×πΏ k = π√πΏ ο³ R = π × π√πΏ + π × πΏ ο³ R = 2,167.61 ο» 2,168 units SS = π × π√πΏ = 2.33*28*√7 = 172.61 ο» 173 units The reorder point for the bar of soap if management want to have a 99% cycle-service level is 2,167 bars of soap. The safety stock is 173 bars of soap, which is the number of units we have to keep in store to face uncertainties. 3.3 π C(Q*) = (π΄ × π∗ + β π∗ 285 2312.98 ) × 365 = (9 × 2312.98 + 9.589π −4 × ( 2 2 + 172.61)) × 365 = $869.95 The total annual cost for the bars of soap, assuming a Q system will be used, is $869.95. 4 Exercise 4 4.2 As we can see in the Excel sheet, there is no change in the production when we only receive 300 units (during week 5 and 6) instead of 600 (during week 5). Indeed, the net requirements during week 5 are lower than 300, therefore we still had an extra 100 units for the week after. Seeing that we already had 120 units in stock at the beginning of week 5, we only needed 80 units to satisfy the net requirements of week 5. However, in that case, for week 6, we would have needed 320 units because the net requirements were 320. 4.3 According to us, the lot-for-lot policies are rather optimized because it allows us to reduce our overall costs. On the one hand, this reduces the inventory holding costs because from one week to another, we do not have any units in our inventory. On the other hand, the fact that we need to order each week might increase the fixed costs but as we do not have any inventory costs, it will not go through the roof and might be less expensive overall. As a consequence, this policy is not as flawed as it may sound. As far as the Fixed-Order-Quantity policies are concerned, we think that we can do a better job to optimize them. In fact, many times, we ordered a tremendous number of units compared to what was actually needed. For instance, the FOQ-policy for the ‘Brackets’ is 600 units. However, this number is too high because we never need such a quantity. On further consideration, we found out that 200 units might be more optimized. Indeed, with this quantity we would be able to order a bigger quantity (such as 600 as initially), as well as a lower quantity (200 or 400 units). Therefore, we think that these new quantities might cut some our costs and be more cost effective. In our opinion, the lot sizing policies for the ‘Oak Sheet’ is not optimal. When the requirements fluctuate a lot from one period to another, the FOQ policy is not a great idea because this forces us to carry stock from week to week; consequently, increasing our holding costs. In the light of our reflection, we think that a L4L policy might be better for this particular product. Seeing that a large number of these units are needed, we must avoid ordering too many items to have as little stock as possible. Switching to this policy will undoubtedly increasing our ordering costs but it will bring our inventory holding costs down. Overall, the lead times are low (one or two weeks) but it is a significant variable to consider. Actually, as we already know what we need for a period, it does not matter when we place the order. 5