Logistic Planning Techniques and Application - Assignment Example

Logistic planning techniques and application Affiliation: Table of Contents TASK 2 Create 2 Assign 3 Process 3 Decide 4 Dispose 4 The Two Replications 5 Needed Changes 6 TASK 2 9 EXPONENTIAL SMOOTHING 9 Functions and Limitations of SES (exponential smoothing) 9 Table Interpretation (see appendix A) 9 Defining the Procedure: Exponential Smoothing 10 Shape on graph 11 Shape on graph 12 Graphs’ Interpretation 12 Trend Line Analysis 13 Graph 1 13 Graph 2 13 MOVING AVERAGE 14 Applications 14 Use 14 Min/Max Points 15 Defining the Procedure: Moving Average 15 Differentiating the Graph Lines 17 APPLICATION OF MINITAB 17 Description 17 Variance in Excel numbers as Compared to Minitab 18 Showing Variance in Original and Forecasted using Minitab 18 References 19 TASK 1 Figure 1: Order Processing using Arena Simulation. Create Figure 1 above shows the process of fixing orders as well as the decision modules involved to ensure that orders are taken with reference to their deadlines (Gahagan, 2008). However, while different orders have different deadlines, the customers fixing these orders choose different packages for having their orders processed for delivery. For instance, two customers are fix two orders, as indicated in the above example. However, these orders do not reach the customers at the same time since the packages for each order is different from the other. This scenario involves two levels of analysis to ensure that the process and the functionality of the system works with reference to priority (Fish, 2001). Assign It is upon the operator, in this case, the packer, to decide which among the two orders should be assigned a higher priority in order to ensure efficient functionality of the system. In the above example, two orders are placed and the two orders are made by different customers who take varying packages for the transportation of their orders from place of origin to the place of delivery. Process After the placement of the two orders, each customer awaits for their order to process with respect to the package they choose. However, one major consideration is the priority of the orders based on the package selected. Thus, the logistics company is able to determine the nature of the order as well as its priority. Rush orders are those orders that have minimal deadline assigned to them based on the urgency of the customer. In this case, the system accounts for rush orders and compares them with the traffic for regular orders which are not considered a priority for the company (Riopel, & Langevin, 2005). One logical argument for the separation of regular orders from rush orders is the fact that rush orders are considered high priority. High priority orders in the logistics industry define the nature of the order as well as the accompanying returns from the order. Since rush orders have a shorter deadline, they have to be accommodated with respect to the returns. However, returns from a rush hour can be compared to returns expected by companies that offer delivery schedules within the same day. Same-day deliveries are deliveries that are considered priority orders that logistics companies either take or decline to offer with respect to lead-time delays, nature of transportation channel and the overall combination of factors that affect the delivery of the orders within the same day. In the current case, the packer is tasked with the analysis of incoming orders and processing them with respect to their priorities. In this case, it is indicated that order 2 is considered a rush order. Decide From the Arena illustration shown above, it is observed that the decisions made by the packer are related to the nature of the order. In this instance, the packer chooses the rush order over the regular order. However, the fact that the packer has the jurisdiction to choose the order that is most beneficial, the exercise undermines the role played by the system as any individual can choose rush orders over regular orders. In this case, the rush and regular orders are required to run automatically based on the fact that priority versus regular orders both contribute to the financial projects of the company. In this case, it is expected that the system must allow some form of balance between rush and regular orders. Dispose Figure 1 above shows that each order coming in passes through the system from which the packer decides which mode of transport suits each type of order. For instance, order 2 is a rush order while order 1 is not. In order for packer to take into consideration the level of urgency of different orders, the modes of transport that suit each order type are reflected. Hence, a same-day delivery situation calls for the accumulation of rush order and assigning them the most suitable mode of transport (Kelton, Sadowski, & Swets, 2009). Hence, the above case calls for transport determination approach where rush orders are either delivered using the most efficient means of transport while regular orders are loaded to transportation means that meet the relaxed customer package. The Two Replications Under the assumption that customer 1 uses the regular package, lower priority is assigned to the order and receiving of the order is expected to be much later than the rush orders. Considering that the same simulation is considered for two replications, the relevance and role played by the system comes handy. Initially, the role of the packer is indicated as decision-making but in actual sense, the packer only assigns order modes of delivery as the rest of the data is automatically simulated. For instance, there are various levels of assigning order priorities. Among these is the logical level and the capacity level. Based on the rate at which regular orders and rush orders are made, the system makes use of rationed processed of rush and regular orders. From two replications, the output of the system is expected to allocate a higher chance of processing a regular order (Melamed, & Altiok, 2010). Assuming that every 3 rush orders make room for 2 regular orders to be processed, it is expected that in the processing of 20 orders, 12 will be rush and 8 will be regular. Thus, if the logistics company operates on a 12-hour shift and the maximum number of order per day is 40 orders, it is expected that rush orders will be grouped with reference to the capacity of the transporting means. For regular orders, the same system would ensure that in a 12-hour duration, 8 orders are processed and the loading capacity of the delivery channel takes 20 orders at a time. This would mean that regular orders will be processed as a function of order interval (2 in every 5), order priority (8 orders in 12 hours), carrier capacity (20 orders at a time), and means of transport (most convenient for the transportation of orders as dictated by the package at hand). Needed Changes Table 1: Replication of a rationed system Graph 1: Rush vs. Regular Orders There are two levels of changes that the company in this case can make. Firstly, figure 1 above does not necessarily show the logical balance between rush and regular orders. For instance, the illustration indicates that Packer picks regular orders over rush orders. However, assuming that there was no ratio of accommodating each type of order, it is the duty of the company to optimize the system such that regular and rush orders are processed using logical simulations such that any rush order has a 65% chance of processing while regular has 35%. Arguing in terms of chance, the system would be flawed as the probabilistic approach would be randomized over all the orders. In this case, the company can determine the processing of the orders with reference to speed of processing. In this case, the company can decide that it will develop a model which processes rush orders at a speed of 65% against the regular speed of 35% (Graph 1 shows the rationed processing of rush to regular orders). For randomly generated based on the ratio of 65:35 for rush to regular orders, the recommended change is represented graphically in graph 1. Hence, for every 100 orders fixed, rush and regular orders will be processed based on the ratio of 13:7 for rush to regular orders. Figure 2: Arena Model for Change 2 Secondly, the company has another option which seeks to optimize output while at the same time balancing between rush and regular orders. The company is not sustained through the delivery of rush orders only and regular customers matter as much as the rush customers do. In this case, the company can separate the system such that rush orders and regular orders are handled by different packers who would determine the mode of transport and the time-scale attached to each order type. Rush orders would be handled differently and regular orders will be delivered with respect to destination as well as the routes ending to the customers’ locations (see figure 2). The current version of the system does not show how destinations affect the effectiveness of handling rush and regular orders within the same system. For instance, if the company is approached by three customers, two for regular orders and one for rush orders, the destinations of these orders must be considered. Hence, the company should determine whether the two regular orders make more if combined as opposed to the amount that comes from the rush orders - see how figure one considers destination to include two levels of decision making. For a company delivering one rush and regular order to a customer in Australia and 1 regular order to the USA, the rush and the regular orders can be grouped if they are headed to the same destination. Thus, the current system must include a decision node between packer and the transportation nodes such that packer labels orders with reference to destinations and priorities. The merit for this approach is to ensure that small rush orders travelling long distances can be coupled with regular orders that are currently leaving for the same location. However, if rush orders’ nature disallows slow means of delivery, orders can be grouped in a manner that would incur the least operational charges to both the customer and the company as well. The Arena Model for the second set of changes is presented below as; TASK 2 EXPONENTIAL SMOOTHING Functions and Limitations of SES (exponential smoothing) The exponential smoothing method is an averaging method for forecasting that makes use of historical data to predict the values of future sales, prices, or volumes of production. Two major advantages of the method is that analysis for forecasts makes use of historical data which most entities have in their databases and the calculation of one instance is all its takes as the rest makes use of the original functions for the first instance. On the other hand, one limitation of the method is that it makes use of historical data; meaning that without this data, exponential smoothing cannot be used. Strengths and Weaknesses As an averaging method, exponential smoothing is a useful method of keeping track of a company or a business model’s performance for specific periods of time. Thus, with this method, one can predict the value of the next period from the historical data. One weakness of the method is that it is appropriate in the prediction of short interval averages like 3-5 month averages for a total duration of 20 months. Large intervals such as the averaging of 200 items from a total of 1000 items requires the very first value to be used. Thus, regardless of how much data is to be used, all historical values must be used in the prediction of the desired final period. The process of feeding the entire data is hectic and time consuming. Table Interpretation (see appendix A) Based on the data in table 2 above, monthly sales orders are groups in category 1 and 2. For a duration of 14 months, the exponential smoothing and moving averages indicate the forecasts of the last month in the data series. Exponential smoothing as indicated in graph 1 below shows the trend line for a series of data (Glen, 2013). The role of the exponential analysis method is to show the actual versus forecasted results within a data series. The actual data series shows the representation of the raw data over a period of time. On the other hand, the trend line that represents the smoothing slope which indicates or predicts the desired value with reference to an increasing or decreasing trend for the forecasted data. Table one shows the numeral and data output from the excel analysis. The graphical representation of the exponential smoothing method is represented as Graphs 1. Defining the Procedure: Exponential Smoothing The procedure of performing an exponential smoothing technic starts with the identification of the primary historical data. This data must comprise of a time line column and the values associated with each time line variable for the number of periods identified for use. From the data (table 1), the exponential smoothing variables are calculated to provide the data for use in the plotting of the associated graphs, graphs 1 and 2. For this case, the columns of table 1, month and sales order1/2 are considered. For column 4, the exponential output data is indicated and labeled as sales orders 1 – the procedure is the same as column 5, sales’ orders 2. In the first cell of the target column, the entry must be equal to the value of first cell of the actual variables’ column. In this case, the smoothing constant, alpha of 0.1, is considered which is needed to calculate the value of the exponential smoothing output for the second cell, B4 in this case. The product of B4 and the alpha is added to the product of E4 and the difference between 1 (0.9) and the smoothing constant, alpha of 0.1. the result for the first output cell can be used to apply the formula to all other cells till the last period that requires forecasting. Graph 1: Exponential Smoothing 1 Shape on graph Graph 1 above shows the lines for actual and forecasted data. Both lines, as color coded in the legend, show linear curve with patterned replications. The graph does not show a mirror curve as the data does not have a decreasing and increasing trends replicating each other. Additionally, the decreases and increases are irregular as they show fluctuations from one observed period to another. Graph 2: Exponential Smoothing Shape on graph Graph 2 above shows the lines for actual and forecasted data. Both lines, as color coded in the legend, show linear curves with quite a visible patterns. The graph shows a mirror curve for the actual data but a rather linear line for the forecast. Unlike in Graph 1, Graph 2 shows repetitive replications taking an overall increasing trend. . Graphs’ Interpretation From Graph 1 above, the actual and the smoothened lines are indicated. From the actual line, the ruggedness of the curve shows that the data incurred major fluctuations. The irregularity of the orders is responsible for the nature of the curve. The same case goes for Graph 2 below. However, the major area for analysis in these graphs is the gradient of the forecast line. The forecast line does not start anywhere on the y-axis since the smoothening line indicates a time series output. In graphs 1 and 2, the forecast line shows the desirable number of orders that would have been made throughout the entire period (Taylor, 2011). The forecast line takes into consideration what other methods such as time-series analysis using regression make use of, the standard deviation. In this case, the forecast line eradicates the need for standard deviation as the process points to a situation where no deviations are allowed. Based on Graph 1 above, it is shown that the number of orders made from Jan13 to the orders made in Mar14 portray a relationship that forecasts Apr14 orders for category 1 to be 63 (rounded to the nearest whole number). Trend Line Analysis Graph 1 The trend line in graph 1 shows the trend for the F, forecast line. In this case, the trend line shows the consistency of the values indicating the line of best fit to normalize the data fluctuations. In this case, the trend line follows the Forecast to indicate that no major errors in the smoothing exercise were incurred. Graph 2 As indicated under Graph 1’s descriptive analysis, the trend of actual orders indicated a rugged curve showing major fluctuations from one month to another. The fluctuations, although may show varying number of peak points, cannot indicate whether the overall orders are in the increase or in the decline. However, using the forecast curve, it is obvious that the number of orders, regardless of the unpredictable fluctuations, have an increasing trend and forecasted data for Apr14 is a product of this smoothening. Hence, unlike in Graph 1, the number of orders fixed in each month accumulate to indicate an increasing trend. MOVING AVERAGE Moving average method is a forecast method that is used in the generation of forecasts for specific time period, n, by finding the mean of the values of the dependent variables. As the periods evolve, values for the new data are added to the average and the observed values from the initial data is subtracted from the averaging resulting to a new average value for the next period of observation. Given there is Simple Moving Average, the formula for finding the forecast for the next period is expressed as SMA = (most recent n Yt values)/n In the above case, Yt is dependent-variables actual value for period t n, is the total number of values considered in the average. Applications The application of the moving average is to show and predict the momentum of an undertaking. With reference to the case at hand, the momentum of sales for orders 1 and 2 average smoothing to determine the potential number of sales and support. One major characteristic of moving average is that it shows that the falling price or demand for an asset stops and reverses the direction at the same level as an essential average. Use The use of moving average is a method for analyzing historical data by taking the averages of defines internal, the period t. The moving average is used to tell the predicted value for a time series data. Thus, with data for 13 months, the moving average produces the value for the 14th month thus telling the momentum of the sales/prices. Min/Max Points Minimum and maximum points of a moving average in a stock chart or graph mean that the sales, prices, or the variables are ready to move up or down for maximum point and minimum points respectively. Thus, a minimum point in the scatter plot, stock chart, or graph representing the moving average indicates less for the dependent variable – the vice versa is true for the maximum point (Devcic, 2014). Defining the Procedure: Moving Average Moving average is a rather easier method than exponential smoothing in that the procedure is easy and time effective. From table 1 above, the values of column 1, column 2, and column 3 are necessary in the calculation of the moving average. For the output, columns 6 and 7 show the moving average output. Using MS Excel, the number of periods considered for averaging, is selected from the total number of periods. This case, three month averages are considered in which, the third cell of column 6 or 7 indicate the moving average values. In this case, three-month intervals are considered for averaging in which the mean-finding procedure for the three selected data sets. The first output is in the third cell of the 6th or 7th column considering sale orders 1 or 2 on table 2. The output for this cell contains the formula for averaging in which dragging the output cell will automatically feed output data to the rest of the cells completing the data set required to plot the necessary graphs. Graph 3: Moving Average – Order 1 Using quarterly basis, the use of moving average indicates the trend of how order fixtures averaged out taking into account specified periods. The outputs, Graphs 3 and 4, show that the average for three month durations indicates major fluctuations that the business model at hand may be unable to decipher whether the graphical output shows increasing or decreasing order fixture trend. However, as indicated in Graph 4, it is useful to use moving averages such that patterns can be identified in terms of how orders are processed (Scmprofrutgers, 2011). Graph 4: Moving Average – Order 2 The moving average output for category-orders indicates the presence of patterns such that an increasing trend with a three-month pattern is replicated. The usefulness of this process is attached to orders affected by seasonal demand (Chaovalitwongse, Furman, & Pardalos, 2009). Differentiating the Graph Lines With reference to the moving average and the exponential smoothing outputs, it is evident that the trend lines for each of the orders in both cases differ regardless of whether same category of data is used. For the exponential smoothing outputs, the forecast values start slightly below the value of 70 and slightly above the value of 120 for graphs 1 and 2. On the other hand, the moving averages start slightly below the value of 62 and visibly above the value of 140. This shows that the trend lines generated under the moving average are remarkably lower than those of the exponential smoothing output (Robb, & Silver, 2002). Following these differences, moving average is considered in this case as the most reliable as the moving average forecasts data slightly higher than moving average. In this case, the entity analyzing these outputs is capable of allocating resources to manufacture items that will meet the orders of the predicted period without deficits (Hatchett, Brorsen, & Anderson, 2010). APPLICATION OF MINITAB Description Minitab is a statistical program created originally for tutoring statistics but has developed to be a general statistics software. With reference to instructional applications, Minitab is considered a comprehensive tool comparable to SPSS, MATLAB and MS Excel in performing most of statistical functions. Variance in Excel numbers as Compared to Minitab EXCEL MINITAB ORDER 1 ORDER 2 ORDER 1 ORDER 2 MAD MAPE MAD MAPE MAD MAPE MAD MAPE 14.27314 423.6086 27.04997 1455.924 11.472 19.599 14.591 9.908 Showing Variance in Original and Forecasted using Minitab References Chaovalitwongse, W., Furman, K., & Pardalos, P. (2009). Optimization and Logistics Challenges in the Enterprise. Springer Science & Business Media Devcic, J. (2014). Simple Moving Averages Make Trends Stand Out. Accessed online on December 6, 2014 from http://www.investopedia.com/articles/technical/052201.asp Fish, G. (2001). Discrete-Event Simulation: Modeling, Programming, and Analysis. Springer Science & Business Media Gahagan, S. (2008). Simulation and Optimization of Production Control for Lean Manufacturing Transition. ProQuest. Glen, S. (2013). How to perform exponential smoothing in Excel 2013. Accessed online on November 28, 2014 from https://www.youtube.com/watch?v=0r5pK8mlcXM Hatchett, R., Brorsen, W., & Anderson, K. (2010). Optimal Length of Moving Average to Forecast Futures Basis. Journal of Agricultural and Resource Economics, Vol. 35, No. 1; pp. 18-33 Kelton, D., Sadowski, R., & Swets, N. (2009). Simulation with Arena. McGraw-Hill Higher Education Melamed, B. & Altiok, T. (2010). Simulation Modeling and Analysis with ARENA. Academic Press. Riopel, D., & Langevin, A. (2005). Logistics Systems: Design and Optimization. Springer Science & Business Media. Robb, D., & Silver, E. (2002). Using Composite Moving Averages to Forecast Sales. The Journal of the Operational Research Society, Vol. 53, No. 11. pp. 1281-1285 Scmprofrutgers. (2011). Moving Average Forecast in Excel. Accessed online on November 28, 2014 from https://www.youtube.com/watch?v=f9qprdj1Er8 Taylor, J. (2011). Multi-item sales forecasting with total and split exponential smoothing. The Journal of the Operational Research Society, Vol. 62, No. 3. APPENDIX A: Excel Data Table Month Sales Order 1 Sales Order 2 Exponential Smoothing Order 1 Exponential Smoothing Orders 2 Moving Average Order 1 Moving Average Order 2 E1 E2 Abs Value 1 Abs Value 2 Err. Sq 1 Err. Sq 2 MAD 1 MAD 2 MSE 1 MSE 2 Jan13 68 124 68 124 68 124 4624 15376 14.27314 27.04997 423.6086 1455.924 Feb13 70 148 68 124 2 24 2 24 4 576 Mar13 56 153 68.2 126.4 -12.2 26.6 12.2 26.6 148.84 707.56 Apr13 48 159 66.98 129.06 64.6666667 141.666667 -18.98 29.94 18.98 29.94 360.2404 896.4036 May13 68 125 65.082 132.054 58 153.333333 2.918 -7.054 2.918 7.054 8.514724 49.75892 Jun13 78 150 65.3738 131.3486 57.3333333 145.666667 12.6262 18.6514 12.6262 18.6514 159.4209 347.8747 Jul13 56 159 66.63642 133.21374 64.6666667 144.666667 -10.6364 25.78626 10.63642 25.78626 113.1334 664.9312 Aug13 50 162 65.572778 135.792366 67.3333333 144.666667 -15.5728 26.20763 15.57278 26.20763 242.5114 686.8401 Sep13 69 130 64.0155002 138.4131294 61.3333333 157 4.9845 -8.41313 4.9845 8.413129 24.84524 70.78075 Oct13 72 158 64.51395018 137.5718165 58.3333333 150.333333 7.48605 20.42818 7.48605 20.42818 56.04094 417.3107 Nov13 50 162 65.26255516 139.6146348 63.6666667 150 -15.2626 22.38537 15.26256 22.38537 232.9456 501.1046 Dec13 67 169 63.73629965 141.8531713 63.6666667 150 3.2637 27.14683 3.2637 27.14683 10.65174 736.9503 Jan14 73 138 64.06266968 144.5678542 63 163 8.93733 -6.56785 8.93733 6.567854 79.87587 43.13671 Feb14 48 160 64.95640271 143.9110688 63.3333333 156.333333 -16.9564 16.08893 16.9564 16.08893 287.5196 258.8537 Mar14 62 168 63.26076244 145.5199619 62.6666667 155.666667 -1.26076 22.48004 1.260762 22.48004 1.589522 505.3521 Apr14 63.1346862 147.7679657 61 155.333333 Note: table is formatted to fit the window. It is however visible in excel by right clicking it and choosing, ‘edit in excel’. Read More