成本與價格的比較.ppt

1、1,製造業物流與供應鏈管理,排程(Ref: Finch 生管chap8) 排程研究模式與解法(Ref: 老師專案心得) 看板後拉、豐田供應鏈管理(Ref: 豐田模式、豐田供應鏈管理) 供應鏈管理(Ref: Chopra供應鏈管理) Forecasting (Ref: Chopra供應鏈管理) Aggregate Planning (簡略)(Ref: Chopra供應鏈管理) 存貨管理(Ref: Finch 生管chap11) 工廠佈置 (簡略),2,排程,3,排程,定義：決定工作完成日期，以及所需的任務和活動 排程可以協助準時完成 直接影響顧客的價值屬性 排程可以影響資源的利用率 直接影響生產

2、率 前推排程 用於起始日期已知，欲決定完成日期的方法 回溯排程 完成日期或截止日期已知，再決定起始日期,4,排程,甘特圖 在一個工作流程中各個步驟的佈置 單一資源中分別進行多個工作,5,傳統工作排序法,排序法則：在單一資源下安排多個工作的方法 先到先服務 最早截止日期(EDD) 工作截止時間最早的優先處理 最短處理時間(SPT) 工作處理時間最短的優先處理 關鍵比例(CR) 計算剩餘時間與完成工作所需時間的比例 比例愈小愈先執行的排序法則 CR 1 表示工作無法準時完成,job1 11 16,job2 11 10,Processing time,Due day,CR1 = (16 - 0) /

3、 11,CR2 = (10 - 0) / 11,目前時間為0,6,範例 8.1：採用 EDD 的工作排序(1/2),1,2,3,4,5,7,範例 8.1：採用 EDD 的工作排序(2/2),=0+3,=3+5,=8+4,=12+11,=23+7,5名顧客,8 / 5,8,(processing time),範例 8.2：採用 SPT 的工作排序,1,2,3,4,5,排序後的結果：,=0+3,=3+4,=7+5,=12+7,=19+11,5名顧客,16 / 5,9,範例 8.3：採用 CR 的工作排序,CR=剩餘工作天數 / 完成工作所需時間,CR,17/11,9/5,6/3,11/4,29/7

4、,=0+11,=11+5,=16+3,=19+4,=23+7,33 / 5,5名顧客,排序後的結果：,Scheduling hybrid flow shops with parallel batch, release time, and machine eligibility constraints,I-Lin Wang*, Yu-Bang Chang* and Taho Yang* *Department of Industrial and Information Management, National Cheng Kung University, Taiwan *Institute of

5、 Manufacturing Engineering National Cheng Kung University, Taiwan,11,Introduction,Background The case in this research a 12-inch semiconductor Fab in Tainan Technology Park, Taiwan Semiconductor Industry The most important industry of Taiwan Competitive and cutting-edge technology Scheduling Improve

6、 efficiency in semiconductor production,12,Problem Definition,There two types of process in manufacturing Serial：If a job contains 25 wafers, a serial process will deal with the wafers one by one till all of them are done, and then proceed to the next machine. Batch：a batch process can deal with 2 t

7、o 6 jobs with the same recipe at the same time. (In this research, the machines at the second stage can manufacture 2 jobs simultaneously.),13,Problem Definition,Serial jobs,Stage-1 machines,Batch jobs,Stage-2 machines,14,Solution Methods,Mixed Integer Liner Programming (MIP) Model Exact optimal sol

8、ution Time consuming Usually not feasible for real-world applications Heuristics Good solution Efficient No guarantee on solution quality,15,Mixed Integer Programming Models,Decision variables in the model,16,Mixed Integer Programming Models,Objective Constrains,constrains on jobs and batches,17,Mix

9、ed Integer Programming Models,constrains,constrains on arrival time and set up time of the second stage,Limit Makespan,constrains on releasing time and finishing time of the two stages,18,Mixed Integer Programming Models,Problem Scale（job=J, machine=M） The number of decision variables： The number of

10、 constrains：,8-2-2 stands for 8 jobs, 2 machines at each stage,16 times of 8-2-2,81 times of 8-2-2,19,Skills to Reduce Variables,Problem scale after reducing variables（job=J, machine=M） The number of decision variables: The number of constrains:,20,Proposed Heuristics,BFIFO This method avoids long w

11、aiting time of batches Combination of BFIFO and MIP model Combination of BFIFO and ITT,time,Method 1,time,Method 2,21,Combination of BFIFO and MIP model,Dispatch rule 2,batch,Mathematic model,first machines,second machines,Research methods,22,Combination of BFIFO and MIP model,Problem scale（job=J, m

12、achine=M） The number of decision variables: The number of constrains:,Research methods,23,Reducing model variables,Problem scale（job=J, machine=M） The number of decision variables: The number of constrains:,Research methods,24,Combination of BFIFO and ITT,Based on the current BFIFO scheduling decisi

13、on, local search techniques that conduct job insertion or interchanges within or between machines can help improve the solution quality very efficiently. Interchange swap jobs on different machines Translocation pick a job from the machine with longest processing time, insert it to another machine T

14、ransposition swap job position on the same machine,Research methods,25,Interchange,Job1,Machine1,Job3,Job5,Job7,Job2,Machine2,Job4,Job6,Job8,Job1,Machine1,Job6,Job5,Job7,Job2,Machine2,Job4,Job3,Job8,Job9,Job10,Job9,Job10,Research methods,26,Translocation Method 1,Job1,Machine1,Job3,Job5,Job7,Job2,Ma

15、chine2,Job4,Job6,Job8,Job1,Machine1,Job5,Job7,Job2,Machine2,Job3,Job4,Job6,Job8,Research methods,27,Translocation Method 2,Job1,Machine1,Job3,Job5,Job7,Job2,Machine2,Job4,Job6,Job8,Job3,Machine2,Job2,Job4,Job6,Job8,Job2,Machine2,Job3,Job4,Job6,Job8,Job2,Machine2,Job4,Job6,Job8,Job3,or,or,Research me

16、thods,28,Comparison of Translocation Methods,Translocation method at the first stage,Research methods,29,Comparison of Translocation Methods,The cause of the difference between two translocation methods,Job1,Machine1,Job3,Job5,Job7,Job2,Machine2,Job4,Job6,Job8,Job1,Machine1,Job5,Job7,Job3,Machine2,J

17、ob2,Job4,Job6,Job8,Job9,Job9,First stage,Research methods,30,Transposition,Job1,Machine1,Job2,Job3,Job4,Job5,Job6,Job2,Machine1,Job1,Job3,Job4,Job5,Job6,Job2,Machine1,Job3,Job1,Job4,Job5,Job6,Job2,Machine1,Job3,Job4,Job5,Job6,Job1,Research methods,31,Computational Experiments and Analyses,Efficiency

18、: running time, Effectiveness: makespan Seven methods: small-scale problemstwo MIP methods (MIP,MIPH)five heuristic methods (FIFO,BFIFO,BFIFO+MIP,BFIFO+MIPH,BFIFO+ITT) Five heuristic methods: large-scale problems Three fastest methods (FIFO,BFIFO,BFIFO+ITT),32,Running time of seven methods,Computati

19、onal Experiments and Analyses,33,Makespan time of seven methods,Computational Experiments and Analyses,34,Running time of five heuristic methods,Computational Experiments and Analyses,35,Makespan time of five heuristic methods,Computational Experiments and Analyses,36,Running time of three fastest m

20、ethods,Computational Experiments and Analyses,37,Makespan time of three fastest methods,Computational Experiments and Analyses,使用較先進的scheduling技巧可能會增加2030%產能,38,Conclusions and Suggestions,MIPH is efficient than MIP, both require much running time for larger cases. Both FIFO an BFIFO can find soluti

21、on quickly, but the solution quality is not acceptable. The combination of BFIFO and ITT can have the best efficiency, and the solutions are satisfactory. We recommend the use of BFIFO+ITT to solve difficult scheduling problems in real-world semiconductor manufacturing.,39,公司的存貨越多 就越不可能達成期望的目標。 豐田生產

22、制度開發者大野耐一,使用後拉制度以避免生產過剩,原則3,40,看板式拉動生產：避免生產過剩,豐田模式不講存貨管理，它講求的是消除存貨。後拉意指即時生產的理想狀態：在顧客(包括生產流程中下一個步驟的內部顧客)需要時，才提供其所需要數量的東西。 豐田生產制度不是零存貨制度，而是依賴應用後拉式制度來補貨的材料貨棧 (stores)。 存貨控管得宜的超級市場。,41,原則：由顧客拉動與進行補貨,由超級市場作業模式，大野耐一認知到，必須有存貨才能維持流程的順利運作。 在各作業之間設立零件或材料的貨棧，以控管存貨，當顧客取用特定項目後，便進行補貨，顧客若未取用某個項目，此項目便停留於貨棧裡，也不會進行補貨

23、。 豐田使用看板制度以管理並確保在即時生產制度下的材料流程與生產。,42,實例：在必要處進行後拉,採行後拉式制度的豐田組裝廠。 組裝廠從經銷商處獲得顧客訂單，生產控管單位擬定穩定均衡的生產時程表。 流程起始於零組件組裝廠，如下圖：,43,存貨導向技術：看板系統,看板(Kanban) 連結生產率與需求的系統 Kanban 即是日語稱為可視的記錄或信號 傳統的生產方式為先預測需求，再批量生產以滿足預測 將產品推向顧客 看板系統是由需求將產品拉往前,44,看板系統於各工作站、各部門和各工廠之間僅使用少量的緩衝存貨，緩衝存貨的最大數量由管理者決定 一旦當緩衝區存貨滿了，即必須停止生產；緩衝區存貨低於最

24、大量時，生產信號即驅動補充緩衝區 當一只零件的容器被生產填裝完成，看板將附在容器上並送至需要的工作站。一旦容器清空，這張卡片將送回授權生產,存貨導向技術：看板系統 (cont.),45,簡單的看板形式：,存貨導向技術：看板系統,46,存貨導向技術：看板系統 (cont.),47,看板的種類： 單卡看板 雙卡看板,存貨導向技術：看板系統 (cont.),48,假設生產線包含 7 個工作站(WCs)，生產線末端是裝有6 個完成品的容器 工作站之間也有一個裝著 6 個零件的容器,緩衝區已滿，故前面的WC未被授權生產,存貨導向技術：看板系統 (cont.),49,當一個完成品從WC7後的緩衝區取走，會

25、發生何事？,存貨導向技術：看板系統 (cont.),50,多出來的空間馬上驅動 WC7 補貨 WC7 需要用 WC6 和 WC7 之間緩衝區的一個零件 只要 WC7 取走一個零件，WC6 後的緩衝區立即驅動生產 WC6 的生產需要前一個緩衝區的一個零件，如此又驅動WC5 的生產。如此由完成品的需求開始啟動，往前驅動通過整個生產線,存貨導向技術：看板系統 (cont.),51,在系統中，WIP 存貨總數為何？ 每一個緩衝區容量為 6 件，故半成品的總和為 42 件 管理者可以改變緩衝區(容器)的大小，來調整半成品數量,存貨導向技術：看板系統 (cont.),52,看板系統也能連結實際上並不相連的

26、工作站 看板不一定是空的容器，也可以卡片、地板上油漆的方格、電子訊息、或其他可表達再做一些！的訊號來替代,存貨導向技術：看板系統 (cont.),53,看板模擬(一),完成品,原物料,工作站 #1,工作站 #2,WIP 存貨,一旦訂單需要一個完成品，工作站2即由WIP存貨中取出一件並開始加工製造，如此亦授權工作站1取用一件原物料開始生產，供應商亦即刻補充原物料,顧客,供應商,原物料,54,完成品,原物料,工作站 #1,工作站 #2,WIP 存貨,一旦工作站2完成生產作業，即將之送至完成品處。工作站1完成其作業，亦將之送至WIP存貨處。此時各緩衝區均達最大容量，故停止生產，直到另一個完成品需求訂

27、單到來,顧客,供應商,原物料,看板模擬(一) (cont.),55,完成品,原物料,工作站 #1,工作站 #2,WIP 存貨,再次，一旦訂單需要一個完成品，工作站2即由WIP存貨中取出一件並開始加工製造，如此亦授權工作站1取用一件原物料開始生產，供應商亦再次即刻補充原物料,顧客,供應商,原物料,看板模擬(一) (cont.),56,完成品,原物料,工作站 #1,工作站 #2,WIP 存貨,再一次，一旦工作站2完成生產作業，即將之送至完成品處。工作站1完成其作業，亦將之送至WIP存貨處。此時各緩衝區均達最大容量，故停止生產,顧客,供應商,原物料,模擬結束,看板模擬(一) (cont.),57,看

28、板模擬(二),58,Storage area,Empty containers,Full containers,Receiving post,Kanban card for product 1,Kanban card for product 2,看板模擬(二) (cont.),59,Storage area,Empty containers,Full containers,Receiving post,Kanban card for product 1,Kanban card for product 2,看板模擬(二) (cont.),60,Storage area,Empty contain

29、ers,Full containers,Receiving post,Kanban card for product 1,Kanban card for product 2,看板模擬(二) (cont.),61,Storage area,Empty containers,Full containers,Receiving post,Kanban card for product 1,Kanban card for product 2,看板模擬(二) (cont.),62,Storage area,Empty containers,Full containers,Receiving post,K

30、anban card for product 1,Kanban card for product 2,看板模擬(二) (cont.),63,Storage area,Empty containers,Full containers,Receiving post,Kanban card for product 1,Kanban card for product 2,Fabrication cell,O1,O2,O3,O2,看板模擬(二) (cont.),64,其中： k = 看板或容器的數量DLT = 在補貨前置時間的平均需求LT = 補貨前置時間S = 安全庫存 C = 每個容器可容納之數量,

31、管理者必須考慮容器或看板的數量，確定系統裡授權生產的存貨數量,存貨導向技術：看板系統 (cont.),65,確定合適的看板數量之計算步驟 確定裝有零件的容器的平均前置時間有幾個小時 生產時間、運輸時間和等待時間的總和 確定在前置時間內下游工作站的每小時生產率對接受零件的需求 以前置時間(小時)乘以生產率(小時)計算而來 增加一份安全庫存量 確定容器的數量 前置時間所需求的零件數量加上安全庫存量，然後再除以一個容器可容納零件數,存貨導向技術：看板系統 (cont.),66,範例：確定合適的看板的數量 生產監督者希望能決定由桿身桿頭組合區到握柄站的流程中，最合適的看板數量，以控管存貨。桿身桿頭組合

32、區處理全部的球桿，而握柄工作站只對須要額外纏繞握柄的球桿加工。其中一些球桿已有橡膠握柄。請使用下列資料計算出需要多少貨架，已知一個貨架可容納 20 支球桿,存貨導向技術：看板系統 (cont.),67,解答： 平均總前置時間 = 生產時間 + 運送時間 + 等待時間 = 0.7 + 0.5 + 1.9 = 3.1小時 在前置時間內的平均需求量 = 前置時間 x 下游工作站之生產率 = 3.1小時 x 35支球桿小時 = 108.5支球桿 增加 10% 的安全存貨 = 108.5 + 10.85 = 119.35支球桿 貨架的數量 = (需求 + 安全存貨)容器大小 = 119.35/20 =

33、5.9675 = 6 個貨架(小數點進位),存貨導向技術：看板系統 (cont.),68,預定時程表的前推式生產制度仍有可用之處,豐田公司內部仍有許多使用依照生產時程表的前推式制度，例如日本把零組件輸出至美國，或美國境內各工廠間轉運零組件時，就是採取這種制度。 當前置期很短時，最適合採行預定時程表的制度，將會是例如每天下零組件訂單，而不是一個月下一次訂單。因此，當必須採行預定時程表制度時，豐田公司盡可能地使前置期縮短。,豐田供應鏈管理,生產排程與作業 (平準化排程範例),70,豐田組裝廠流程,鋼捲,鋼片存放區,待修正成車存放區,71,生產排程,以銷售訂單及預測為依據，每月制定一次 銷售部門每月

34、會呈交一份三個月訂單及預測 生產控制部門需制訂每日生產計劃,Input： 本月確定訂單，未來兩個月的預測訂單,Output： 本月每日生產計劃,72,排程的輸入資料,接下來要排程的生產月份N確定的訂單保證，與N+1、N+2月的預測訂單 生產月曆與工廠的作業計畫 Ex.休假時程、計畫性加班 限制 Ex.某條組裝線僅能安裝某些引擎型式,73,排程的流程,使用銷售訂單與預測資料，為每輛車建立個別記錄與一獨特的查詢號碼 平準化流程安排製造車子的時程 制定一個均衡或平均的生產計劃,74,生產順序,取決於工廠的作業條件 塗料的批量 勞力密集相關的配備 配備的平準化 用來制定零件訂單，同時做為組裝線控制統的

35、輸入資料,75,生產計劃例子,假設每輛車只有三項配備 等級: LE、XLE 引擎: 4缸、6缸 顏色: 紅、黑、藍 生產優先考量為等級，其次是引擎 為十輛車制定五天的生產時程,76,十輛車之規格,77,每日時程表,每天組裝線上組裝車的順序,順序,78,將相似之規格排在一起,相近規格的車應分散在不同天 目的：平準化流程,79,排程,交換,80,最終排程結果,81,平準化之重要(1/2),工廠能以有效率、有成效之方法運作 維持整條供應鏈的穩定,代工廠,第一層 供應商,下訂單,第二層 供應商,下訂單,產能 1000,產能 5000,第一天訂單 數量1000,第二天訂單 數量500,第三天訂單 數量2

36、500,訂單數量4000,訂單數量2000,訂單數量10000,數量超過各供應商產能!,第一層供應商缺1500,第二層供應商缺5000,若採平準化生產，每天生產1000,第一層供應商缺 2500-(1000*3-1000-500)=1000,第二層供應商缺10000-(4000*3-4000-2000)=4000,三天的產能扣除第一天和第二天的需求量,82,平準化之重要(2/2),降低庫存或貨物短缺的風險 維持完美的平準化是不可能的，變化介於5%就算是好的平準化結果 Ex.平均訂單數量是1000，其範圍可介於950 1050間,83,反思重點,利用平準化來規劃與分配各時期的多樣性，以平衡工作

37、利用以速率為基礎的規劃來平衡供應鏈的流動，從而保持速度 以平準化來穩定工作量及負擔，從而縮減變異性 藉由消除庫存、簡化規劃、建立認同等，可以建立能見度,84,供應鏈管理,85,物流？運籌？ Logistics？,Physical Distribution之日語翻譯（狹義 ）,“物流是供應鏈程序的一部分，及專注於物品、服務及相關資訊 ，從起源點到消費點之有效流通及儲存的企劃、執行與控管， 以達到顧客的要求。” （美國物流管理協會 ）,“整合相關的資訊流、商流、物流、資金流等，使生產的整個 後勤支援與供應鏈管理達到最佳化 。” （運籌管理 ）,物流原料物流生產物流銷售物流廢棄物物流,淺論物流,86

38、,物流運籌的領域,淺論物流,87,物流中心地點選擇,客戶分布 零售店，人口密集的都會區 供應商分布 安全庫存可較低 國內一般進貨的輸送成本由供應商負擔 交通條件 交通方便、多樣交通工具、運送時間 土地條件 工業區、住宅區、商業區、農業區、倉儲區、物流專業園區 價格、法規 環境條件 濕度、鹽度、震動 人力資源條件 24hr? 人力、交通、薪資水準 行政條件 企業優待、土地提供、減稅、都市計畫、土地開發、道路建設、產業政策,物流中心,88,物流中心據點數評估,以最低成本為考量 物流總體成本輸配送成本+據點管理成本+庫存成本總和 -據點數少 多 ,物流中心,89,Locating Plants re

39、ly on judgment and opinion (good for cases with little historical data, or when experts have market intelligence) Time Series: use historical demand only Static revise estimates for level and trend for period 1 and for seasonal factor for period 5 L1 = a(D1/S1)+(1-a)(L0+T0) = (0.1)(8000/0.47)+(0.9)(

40、18439+524)=18769 T1 = b(L1-L0)+(1-b)T0 = (0.2)(18769-18439)+(0.8)(524) = 485 S5 = g(D1/L1)+(1-g)S1 = (0.1)(8000/18769)+(0.9)(0.47) = 0.47 F2 = (L1+T1)S2 = (18769 + 485)(0.68) = 13093 If no info indicating whether there is a trend or seasonality, what to do? Check forecast error.,Trend- and Seasonali

41、ty-Corrected Exponential Smoothing (continued),125,Measures of Forecast Error,Help to determine whether the current forecasting method is accurately predicting the systematic component of demande.g. always postive error overpredicting systematic component Help manager to consider contingency plane.g

42、. mail order with 2 suppliers: one in Far East with 2month lead time; the other one is a local supplier but more expensive; Forecast error = Et = Ft - Dt Mean squared error (MSE) (relate to variance of the FE) MSEn = (Sum(t=1 to n)Et2)/n Absolute deviation = At = |Et| Mean absolute deviation (MAD) (

43、estimate the standard deviation of the random component, assuming which is normally distributed with mean 0) MADn = (Sum(t=1 to n)At)/n s = 1.25MAD,126,Measures of Forecast Error,Mean absolute percentage error (MAPE) MAPEn = (Sum(t=1 to n)|Et/ Dt|100)/n Bias Shows whether the forecast consistently u

44、nder- or overestimates demand; should fluctuate around 0 biasn = Sum(t=1 to n)Et Tracking signal should be within the range of +6, Otherwise (i.e. underforecasting: TS6), possibly use a new forecasting method TSt = bias / MADt,127,Forecasting Demand at Tahoe Salt,Moving average Simple exponential sm

45、oothing Trend-corrected exponential smoothing Trend- and seasonality-corrected exponential smoothing Table 7.2 on page 199 shows Winters model best fits,128,Forecasting in Practice,Collaborate in building forecasts The value of data depends on where you are in the supply chain Reams and reams of dat

46、a need not to be shared across the SC E.g. POS is useful for retailer, but not for manufacturer; aggregate demand data from retailer is valuable for manufacturer, but not for retailer Think what data is valuable to each member and share only those data Be sure to distinguish between demand and sales

47、 Notice that unmet demand due to stock-outs, competitor actions, pricing, promotionsetc Not easy to adjust, but if done, will increase accuracy and SC performance,129,Aggregate Planning,130,Role of Aggregate Planning in a Supply Chain,Main objective: to specify operational parameters over the time h

48、orizon: Production rate: units per unit time Workforce: workers/capacity needed Overtime: overtime production planned Machine capacity level: machine capacity needed Subcontracting: subcontracted capacity required Backlog: demand unsatisfied Inventory on hand: inventory planned Allows the SC to alte

49、r capacity allocation and change supply contracts All supply chain stages should work together on an aggregate plan that will optimize supply chain performance,131,The Aggregate Planning Problem,Given the demand forecast for each period in the planning horizon, determine the production level, invent

50、ory level, and the capacity level for each period that maximizes the firms (supply chains) profit over the planning horizon Specify the planning horizon (typically 3-18 months) Specify the duration of each period Specify key information required to develop an aggregate plan,132,Key Information Neede

51、d foran Aggregate Plan,Demand forecast in each period Production costs labor costs, regular time ($/hr) and overtime ($/hr) subcontracting costs ($/hr or $/unit) cost of changing capacity: hiring or layoff ($/worker) and cost of adding or reducing machine capacity ($/machine) Labor/machine hours req

52、uired per unit Inventory holding cost ($/unit/period) Stockout or backlog cost ($/unit/period) Constraints: limits on overtime, layoffs, capital available, stockouts and backlogs,133,Outputs of Aggregate Plan,Production quantity from regular time, overtime, and subcontracted time: to determine numbe

53、r of workers and supplier purchase levels Inventory held: to determine warehouse space and working capital Backlog/stockout quantity: to determine customer service levels Workforce hired/laid off: to determine any labor issues that will be encountered Machine capacity increase/decrease: to determine

54、 if new production equipment needs to be purchased A poor aggregate plan can result in lost sales, lost profits, excess inventory, or excess capacity,134,Aggregate Planning Strategies,Trade-off between capacity (regular time, overtime, subcontracted), inventory, backlog/lost sales Chase strategy usi

55、ng capacity as the lever Time flexibility from workforce or capacity strategy using utilization as the lever Level strategy using inventory as the lever Mixed strategy a combination of one or more of the first three strategies,135,Chase Strategy,Production rate is synchronized with demand by varying

56、 machine capacity or hiring and laying off workers as the demand rate varies However, in practice, it is often difficult to vary capacity and workforce on short notice Expensive if cost of varying capacity is high Negative effect on workforce morale Results in low levels of inventory Should be used

57、when inventory holding costs are high and costs of changing capacity are low,136,Time Flexibility Strategy,Can be used if there is excess machine capacity Workforce is kept stable, but the number of hours worked is varied over time to synchronize production and demand Can use overtime or a flexible

58、work schedule Requires flexible workforce, but avoids morale problems of the chase strategy Low levels of inventory, lower utilization Should be used when inventory holding costs are high and capacity is relatively inexpensive,137,Level Strategy,Maintain stable machine capacity and workforce levels with a constant output rate Production is not synchronized with demand Shortages and surpluses result in f