CASE STUDY: JPMS CHEMICALS CASE STUDY

JPMS Case Study

After reviewing the JPMS case study on pages 277–288 of Supply Chain Network Design, prepare an executive report that summarizes the following dimensions of the scenarios developed in the chapter:

  • Support for growth of the Indian chemical market.
  • Financial results, including profit and capital expenditures—assume the cost of a new plant is $200 million (about 13 billion rupees).
  • Customer service.
  • Operational complexity.

In your discussion of the dimensions, include the following points:

  • Evaluate the ability for the supply chain network to scale for changing market conditions. Provide effective strategies to scale the operations.
  • Evaluate the financial implications of the various scenarios in the supply chain network design. Make recommendations to improve financial results under the different scenarios.
  • Evaluate the customer service implications of various scenarios in the supply chain network design. Make recommendations to optimize customer service.
  • Evaluate the operational complexity implications of the various scenarios in the supply chain network design. Make recommendations for other scenarios that would manage these complexities and also meet other goals.
  • Make sure you use the articles you have read in this course and the text materials to develop a 8+ page paper analyzing these issues.

 

  1. Below is the Case and the Grading Criteria

 

 

 

 

 

 

 

 

 

 

CASE STUDY: JPMS CHEMICALS CASE STUDY

In this chapter, we will look at modeling and analyzing capacity from a manufacturing perspective, as well as how manufacturing and sourcing optimization can be modeled. We will try various what-if scenarios to explore different network alternatives with an example of a chemical company in India.

Indian Chemical Company—Case Study

In this case study, we will review the case of JPMS Chemicals Pvt. Ltd., a large manufacturer of various chemicals. JPMS Chemicals is a privately owned business based in Mumbai, with sales and distribution to other chemical manufacturers and distributors all over India. The company operates a network of four plants located in Madurai (Tamil Nadu), Aurangabad (Maharashtra), Dewas (Madhya Pradesh), and Lucknow (Uttar Pradesh).

The company manufactures and distributes several types of products, including resins and solvents. For the purposes of this modeling exercise, we will simplify them into three product families: Chemical A, Chemical B, and Chemical C. Note that these are the aggregated products that we are modeling—each family includes hundreds of SKUs with similar characteristics.

The map in Figure 15.1 shows the current four plants overlaid against their customer demand points that they ship to. Their customer base includes small distributors as well as manufacturing plants or warehouses, and this aligns with overall population demographics.

 

Figure 15.1 Map of Customers and Plant Locations

All three products can be manufactured at all the plants, so they serve as sources of regional supply—that is, each plant produces products to serve its regional customer base. The table in Figure 15.2 shows the maximum production capacity (in units) available by plant by product family. For Chemical A, the maximum output for one production line is 550,000 units, and both Aurangabad and Lucknow plants have one line each for Chemical A. The other two plants have two lines for Chemical A. For Chemical B, all plants except Aurangabad have one line only. For Chemical C, all plants except Dewas have two lines each (capacity of 11,000 units/line), and the Dewas plant has three lines dedicated to Chemical C.

Figure 15.2 Production Capacity by Product by Plant

With the ongoing economic boom in India, the management team was projecting aggressive growth in demand over the next few years and was concerned whether there was sufficient manufacturing capacity to support growth. They wanted to understand whether they needed a new plant, and if so where and how big it should be.

When we analyze manufacturing or sourcing strategies, it is important to factor the sourcing of raw materials and components into these plants—that is, we want to consider the impact of inbound transportation and sourcing costs. In this case, the components used to manufacture these chemicals are widely available and can be locally procured even at any new locations at roughly the same cost. Based on this, we are not including raw materials and components in this case study.

Let’s start by building a baseline model of the current network with the last 12 months of demand. It always makes sense to start with a baseline model to ensure that the model is accurate and is a good representation of how the supply chain operated.

From Figure 15.3, we can see that the customers are regionally sourced, with some overlap, from the four plants in the network. This has traditionally been the company’s sourcing strategy because this allowed them to reduce transportation costs and provide good customer service.

Figure 15.3 Map Showing Results of Baseline Model

The chart in Figure 15.4 shows that three of the four plants are highly utilized without much spare capacity available. There is some capacity available in the Madurai plant, but given this plant is located in the southernmost part of the country, it will be very expensive to ship across the country when other plants are out of capacity. This shows that there will likely be a need for additional capacity to support the aggressive growth plans.

Figure 15.4 Capacity Utilization by Plant by Product

The JPMS management team was estimating the business to grow by 20% within the next 3 years, and wanted to understand what this would mean in terms of available manufacturing capacity.

Let’s run a what-if scenario with the current plants and capacity but with a 20% increase in customer demand. We will run this with the overall objective set to maximize profits, rather than minimize costs, in order to allow the model to develop a feasible solution. In a model with maximum profits as the objective, we no longer have to enforce that all demand is met. Because meeting each demand adds to the overall profit (if the revenue is high enough), the model will naturally want to meet all the demand. In this case, we are trying to understand how much demand is unmet. In reality, JPMS wants to meet all the demand. Note: This is a common modeling trick used to troubleshoot a minimum cost model that cannot meet all the demand.

After running the scenario, let’s see how the results compared when looking at costs and profits because we ran the growth scenario in profit-maximization mode (see Figure 15.5).

Figure 15.5 Scenario Comparison—Baseline Demand Versus 20% Demand Growth (Current Network)

The cost figures in Figure 15.5 give us some indication of what the model is doing given the increased demand growth. Firstly, the production cost increased by 20% compared to the baseline. Because we did not change the unit production costs, the increase in production costs can be attributed only to the higher volume (20%) being produced. This also tells us that the current network had sufficient capacity to meet the 20% demand growth. Secondly, we notice that the transportation costs increased by 28% compared to the baseline. We had held the transportation rates constant, so this is not a contributor to the higher freight costs. Also, we had applied the 20% growth uniformly across all customer demand points—so there are no variations in growth demographics that would lead to a disproportionately high increase in transportation costs. This tells us that the larger increase in transportation is due to longer transportation distances compared to the baseline, which could be generated by assigning customer demand to plants where spare capacity is available.

A quick look at the maps for the two scenarios illustrates the capacity situation (see Figure 15.6). In the baseline, we can see that the Madurai plant (South India) serves demand only in South India. This was also the plant with some spare capacity available. When the demand was increased, we see that Madurai is now serving customers in East (Orissa) and Central India.

Figure 15.6 Map Comparison—Baseline Demand Versus 20% Demand Growth

When we look at the production capacity utilization by plant and compare the numbers between baseline and the demand growth scenario (see Figure 15.7), this confirms what we see in other reports. We see that the Madurai plant is close to 100% utilized across all three products, whereas Dewas has some spare capacity for Chemical A.

Figure 15.7 Production Capacity Utilization—Baseline Versus Scenario with 20% Growth

When we look at the customer service aspect of this scenario (see Figure 15.8), this shows another key aspect that is important to consider when evaluating new alternatives.

 

Figure 15.8 Services Levels (% of Demand Served) by Distance Band

This shows that service levels were similar for the 800km distance bands, but started deteriorating for longer distances. This makes sense as the model would try to serve demand within short distances first, and serve remaining demand from wherever capacity is available.

The preceding scenario tells us and the management team a few things:

  • ■ There is capacity available within the current network to support 20% demand growth.
  • ■ However, this will mean higher logistics costs in order to utilize available capacity. This would also have a negative impact on customer service levels, especially for customers located beyond 800km from plants.
  • ■ This also tells us that there is very limited spare capacity available to handle demand spikes higher than 20%.

Given the capacity sensitivity, the management team wanted to understand the best location(s) to open a new plant. To answer this question, we will add a list of potential plant locations based on customer and demand demographics. For the purpose of this analysis, we will add a list of 30 locations to the model. The list of locations and their placement on a map is shown in Figure 15.9.

Figure 15.9 List of Potential Plant Locations

For the new plant(s), there are three main questions we are trying to answer:

  • Where should the new plant be located?
  • What products should be manufactured at the new plant(s)?
  • How much capacity needs to be added at the new location(s)—that is, how many production lines do we need to add?

In terms of production capability, we want the model to determine what to make at the new plant(s), so we will give the potential plants the ability to manufacture all products.

As far as capacity is concerned, there are a couple of ways to model this:

  • We could give the new plants unlimited capacity to produce any products. This will allow the model to determine how much product should be produced—we can then determine how many lines are needed based on this.
  • We could add capacity representing a certain number of lines, and let the model utilize this available capacity.

Both options are valid and applicable to model capacity. The best approach will depend on the nature of the business, capital equipment costs, and so on.

The advantage of Option 1 is that it would give the model the most flexibility in determining product sourcing volume from the new plant. Alternatively, this could also be viewed as a disadvantage because the model may come up with a volume mix that represents 1.05 lines—that is, 5% capacity utilized for a second line. It will probably not make sense to open a second production line to handle a very small incremental spillover volume.

Given this, we will use Option 2 and model a specific number of lines at each potential plant and see how the model utilizes this capacity. To give the model sufficient flexibility, we will give each new plant three potential lines for Chemicals A and B, and four potential lines for Chemical C. This allows the new plants to be potentially much larger than the existing plants.

For now, we will not add any opening costs or fixed costs for the new lines—we will look at this a bit later. As we have seen in earlier chapters, we can evaluate the impact of fixed costs without having to model them directly in the model.

Figure 15.10 shows the existing and potential plants with the available lines and capacity.

Figure 15.10 Production Capacity by Plant for Current and Potential Plants

We will also need to add transportation rates and lanes from the new plants (and attached warehouses) to all the customers. We will also create a plant site grouping for all potential plants, allowing us to apply constraints on the number of new plants that can be opened.

We are now ready to run our new set of scenarios with the new plant locations. The first scenario we will run would be to allow the model to select up to one new plant location.

From the map shown in Figure 15.11, we can see that the model has opened a new plant in Dhanbad (Jharkhand) that is serving customers in the eastern part of the country, and the Lucknow plant is serving customers in the northern region only.

Figure 15.11 Map Showing Results of Scenario with One New Plant

When we look at the costs comparison (see Figure 15.12) against the scenario without new plants, we see that the overall costs reduced by 5%—the savings are attributed primarily in outbound transportation costs because the model now has sufficient capacity to serve demand regionally.

Figure 15.12 Cost Comparison for Scenarios With and Without New Plant

The production capacity analysis (see Figure 15.13) shows that at the new plant the model opened two production lines each for Chemicals A and B, and three lines for Chemical C. We can see that three of the seven lines opened are highly underutilized. This is because the model did not have any penalties to open a new line, so it opened a new line if it was cheaper to serve even only a handful of customers from the new plant. We know that it would not make sense to open a new line that is only 5% to 10% utilized. We could easily address this problem by adding a fixed opening cost for each line.

Figure 15.13 Comparison of Production Capacity Utilization Between Scenarios

We will add a fixed cost of Rs 5 million for each line based on the average cost for a new line.

The results of this scenario are shown in Figures 15.14 and 15.15. The chart in Figure 15.14 shows that only one line is opened in the new Dhanbad plant when we add a fixed line cost. The volume served by Dhanbad in the previous scenario is now picked up by Aurangabad and Dewas plants as we see their capacity utilization going up significantly. The consequence of this shift is seen in the cost comparison table. We see that the outbound transportation cost goes up by almost Rs 5 million when we add the fixed line cost at Dhanbad.

Figure 15.14 Production Capacity Utilization Comparison—New Plant With and Without Fixed Line Costs

Figure 15.15 Scenario Results Comparison—New Plant With and Without Fixed Line Costs

Note that comparing these two scenarios is not an “apples-to-apples” comparison because they do not have the same cost structure. However, we are doing this to help gain insight on the impact of adding a new cost component on the solution as previously developed. The management team can use this insight to understand the key drivers in their supply chain, and thereby make more informed decisions.

 

Single-Sourcing

The management team had received feedback from some of their customers that they would prefer to receive all their shipments from one plant only. As we can see in the baseline map, there are several customers that are served from more than one plant. This multisource customer assignment was also turning out to be a challenge operationally for the logistics department. So the management team was interested in understanding the impact of single-sourcing individual customer locations.

Let’s run a new scenario allowing an additional plant but with the constraint that customers must be single sourced.

The map in Figure 15.16 shows that the new plant in Dhanbad is still valid. The scenario costs (see Figure 15.17) show that there is a less than 1% difference in total costs due to the single-sourcing constraint. This shows that the new network is capable of handling customer single-sourcing with relatively no impact on logistics costs. This has a very positive impact on operational complexity because each customer can be assigned to a plant for order processing and shipping.

Figure 15.16 Map Showing Results of Scenario with Customers Single-Sourced

Figure 15.17 Results Comparison of Scenarios With and Without Customer Single-Sourcing

State-Based Single-Sourcing

To evaluate certain tax policies and their implications, the management team wanted to understand the impact of serving each individual state from one plant only—that is, single-sourcing of all customer demand within each state.

To do this, we have created customer groupings by state and applied a constraint in the model that forces each customer group (state) to be sourced from only one plant. The map in Figure 15.18 shows the results of the model with each state single-sourced from one plant/warehouse only. First of all, we notice that the new plant location is not Dhanbad anymore—it has moved east to Howrah instead. Also, we see that the customer assignments look a bit odd—the Madurai plant takes over serving some customers much further North than its previous territory, as well as the new Howrah plant serving customers all the way in the Northwest.

Figure 15.18 Map Showing Results of Scenario with States Single-Sourced

Why is this happening? As discussed in Chapter 5, “Adding Capacity to the Model,” we are creating a harder knapsack problem when we aggregate demand into larger chunks. In this case, the model is forced to look for a solution in which each plant supplies the entire demand of the states it serves. Because capacity is tight, we may get answers that look strange.

Even worse than a map that looks strange, we can see from Figure 15.19 that with this constraint JPMS can only meet 98% of the demand. The capacity constraints combined with the constraint to serve a state from one plant do not allow the flexibility to meet all demand. This analysis gives the management team at JPMS some idea of what kind of impact this constraint may have on their supply chain.

Figure 15.19 Cost Comparison of Scenario with States Single-Sourced

A quick look at the production capacities by plant by line (see Figure 15.20) shows a picture that is a bit hard to understand without reminding ourselves about the sourcing constraints. The scenario on top represents the previous solution with the new Dhanbad plant. Each plant is highly utilized relative to demand in its region, and one new line opened at Dhanbad. When we ran the scenario with each state being single-sourced, we see that some of the lines are fully utilized whereas others are hardly used. Why is this happening? First of all, the demand distribution across the states is not similar for all products—that is, some states have higher demand for Chemicals A and B and very little demand for C. The model is factoring this constraint, along with transportation costs, and fixed line costs to come up with the capacity utilization profile.

Figure 15.20 Production Capacity Comparison of Scenarios With and Without Sourcing Constraints

To reiterate our point from earlier, we see that this single-sourcing constraint added a high level of complexity to the model. At the same time, it also provides us with an opportunity to understand the dynamics of the supply chain and how different variables impact each other.

Based on the multiple scenarios run so far, as well as others run to test other potential plant locations, the management team had sufficient information needed to make a sound decision on their manufacturing strategy. Note that we have focused primarily on key supply chain costs that were considered part of the network model—there are several other qualitative factors that impact the choice of a new plant, such as tax benefits, proximity to key customers, availability of skilled labor, and so on.

The management team will need to consider these factors before they make a decision; however, they can combine these non-quantifiable factors with good cost and service data to understand the value of these the non-quantifiable factors.

Lessons Learned from the Case

We developed a model focused on manufacturing capacity and analyzing the best location for a new plant. We could see that capacity constraints were important to understand the key questions that we were trying to answer. However, the capacity constraints can also make the model far more complex, thereby yielding results that may not be intuitive or make sense. This is especially applicable when we start applying sourcing constraints at the aggregate level (for example, by single sourcing a state). In addition, we also reviewed the importance of modeling fixed line costs and how this impacts decisions regarding how many new lines are needed to support demand.

Grading Criteria

 

Criteria Non-performance Basic Proficient Distinguished
Evaluate a supply chain network’s ability to scale for changing market conditions.
20%
Does not summarize a supply chain network’s ability to scale for changing market conditions. Summarizes a supply chain network’s ability to scale for changing market conditions. Evaluates a supply chain network’s ability to scale for changing market conditions. Evaluates a supply chain network’s ability to scale for changing market conditions; provides effective strategies to scale operations.
Evaluate the financial implications of various scenarios in supply chain network designs.
20%
Does not summarize the financial implications of various scenarios in supply chain network designs. Summarizes the financial implications of various scenarios in supply chain network designs. Evaluates the financial implications of various scenarios in supply chain network designs. Evaluates the financial implications of various scenarios in supply chain network designs; recommends actions to improve financial results under different scenarios.
Evaluate the customer service implications of various scenarios in supply chain network design.
20%
Does not summarize the customer service implications of various scenarios in supply chain network design. Summarizes the customer service implications of various scenarios in supply chain network design. Evaluates the customer service implications of various scenarios in supply chain network design. Evaluates the customer service implications of various scenarios in supply chain network design; recommends actions to optimize service.
Evaluate the operational complexity implications of various scenarios in supply chain network design.
20%
Does not summarize the operational complexity implications of various scenarios in supply chain network design. Summarizes the operational complexity implications of various scenarios in supply chain network design. Evaluates the operational complexity implications of various scenarios in supply chain network design. Evaluates the operational complexity implications of various scenarios in supply chain network design; recommends scenarios that manage complexity while meeting other goals.
Write in a scholarly manner by providing validation and scholar

JPMS Case Study

After reviewing the JPMS case study on pages 277–288 of Supply Chain Network Design, prepare an executive report that summarizes the following dimensions of the scenarios developed in the chapter:

  • Support for growth of the Indian chemical market.
  • Financial results, including profit and capital expenditures—assume the cost of a new plant is $200 million (about 13 billion rupees).
  • Customer service.
  • Operational complexity.

In your discussion of the dimensions, include the following points:

  • Evaluate the ability for the supply chain network to scale for changing market conditions. Provide effective strategies to scale the operations.
  • Evaluate the financial implications of the various scenarios in the supply chain network design. Make recommendations to improve financial results under the different scenarios.
  • Evaluate the customer service implications of various scenarios in the supply chain network design. Make recommendations to optimize customer service.
  • Evaluate the operational complexity implications of the various scenarios in the supply chain network design. Make recommendations for other scenarios that would manage these complexities and also meet other goals.
  • Make sure you use the articles you have read in this course and the text materials to develop a 8+ page paper analyzing these issues.

 

  1. Below is the Case and the Grading Criteria

 

 

 

 

 

 

 

 

 

 

CASE STUDY: JPMS CHEMICALS CASE STUDY

In this chapter, we will look at modeling and analyzing capacity from a manufacturing perspective, as well as how manufacturing and sourcing optimization can be modeled. We will try various what-if scenarios to explore different network alternatives with an example of a chemical company in India.

Indian Chemical Company—Case Study

In this case study, we will review the case of JPMS Chemicals Pvt. Ltd., a large manufacturer of various chemicals. JPMS Chemicals is a privately owned business based in Mumbai, with sales and distribution to other chemical manufacturers and distributors all over India. The company operates a network of four plants located in Madurai (Tamil Nadu), Aurangabad (Maharashtra), Dewas (Madhya Pradesh), and Lucknow (Uttar Pradesh).

The company manufactures and distributes several types of products, including resins and solvents. For the purposes of this modeling exercise, we will simplify them into three product families: Chemical A, Chemical B, and Chemical C. Note that these are the aggregated products that we are modeling—each family includes hundreds of SKUs with similar characteristics.

The map in Figure 15.1 shows the current four plants overlaid against their customer demand points that they ship to. Their customer base includes small distributors as well as manufacturing plants or warehouses, and this aligns with overall population demographics.

 

Figure 15.1 Map of Customers and Plant Locations

All three products can be manufactured at all the plants, so they serve as sources of regional supply—that is, each plant produces products to serve its regional customer base. The table in Figure 15.2 shows the maximum production capacity (in units) available by plant by product family. For Chemical A, the maximum output for one production line is 550,000 units, and both Aurangabad and Lucknow plants have one line each for Chemical A. The other two plants have two lines for Chemical A. For Chemical B, all plants except Aurangabad have one line only. For Chemical C, all plants except Dewas have two lines each (capacity of 11,000 units/line), and the Dewas plant has three lines dedicated to Chemical C.

Figure 15.2 Production Capacity by Product by Plant

With the ongoing economic boom in India, the management team was projecting aggressive growth in demand over the next few years and was concerned whether there was sufficient manufacturing capacity to support growth. They wanted to understand whether they needed a new plant, and if so where and how big it should be.

When we analyze manufacturing or sourcing strategies, it is important to factor the sourcing of raw materials and components into these plants—that is, we want to consider the impact of inbound transportation and sourcing costs. In this case, the components used to manufacture these chemicals are widely available and can be locally procured even at any new locations at roughly the same cost. Based on this, we are not including raw materials and components in this case study.

Let’s start by building a baseline model of the current network with the last 12 months of demand. It always makes sense to start with a baseline model to ensure that the model is accurate and is a good representation of how the supply chain operated.

From Figure 15.3, we can see that the customers are regionally sourced, with some overlap, from the four plants in the network. This has traditionally been the company’s sourcing strategy because this allowed them to reduce transportation costs and provide good customer service.

Figure 15.3 Map Showing Results of Baseline Model

The chart in Figure 15.4 shows that three of the four plants are highly utilized without much spare capacity available. There is some capacity available in the Madurai plant, but given this plant is located in the southernmost part of the country, it will be very expensive to ship across the country when other plants are out of capacity. This shows that there will likely be a need for additional capacity to support the aggressive growth plans.

Figure 15.4 Capacity Utilization by Plant by Product

The JPMS management team was estimating the business to grow by 20% within the next 3 years, and wanted to understand what this would mean in terms of available manufacturing capacity.

Let’s run a what-if scenario with the current plants and capacity but with a 20% increase in customer demand. We will run this with the overall objective set to maximize profits, rather than minimize costs, in order to allow the model to develop a feasible solution. In a model with maximum profits as the objective, we no longer have to enforce that all demand is met. Because meeting each demand adds to the overall profit (if the revenue is high enough), the model will naturally want to meet all the demand. In this case, we are trying to understand how much demand is unmet. In reality, JPMS wants to meet all the demand. Note: This is a common modeling trick used to troubleshoot a minimum cost model that cannot meet all the demand.

After running the scenario, let’s see how the results compared when looking at costs and profits because we ran the growth scenario in profit-maximization mode (see Figure 15.5).

Figure 15.5 Scenario Comparison—Baseline Demand Versus 20% Demand Growth (Current Network)

The cost figures in Figure 15.5 give us some indication of what the model is doing given the increased demand growth. Firstly, the production cost increased by 20% compared to the baseline. Because we did not change the unit production costs, the increase in production costs can be attributed only to the higher volume (20%) being produced. This also tells us that the current network had sufficient capacity to meet the 20% demand growth. Secondly, we notice that the transportation costs increased by 28% compared to the baseline. We had held the transportation rates constant, so this is not a contributor to the higher freight costs. Also, we had applied the 20% growth uniformly across all customer demand points—so there are no variations in growth demographics that would lead to a disproportionately high increase in transportation costs. This tells us that the larger increase in transportation is due to longer transportation distances compared to the baseline, which could be generated by assigning customer demand to plants where spare capacity is available.

A quick look at the maps for the two scenarios illustrates the capacity situation (see Figure 15.6). In the baseline, we can see that the Madurai plant (South India) serves demand only in South India. This was also the plant with some spare capacity available. When the demand was increased, we see that Madurai is now serving customers in East (Orissa) and Central India.

Figure 15.6 Map Comparison—Baseline Demand Versus 20% Demand Growth

When we look at the production capacity utilization by plant and compare the numbers between baseline and the demand growth scenario (see Figure 15.7), this confirms what we see in other reports. We see that the Madurai plant is close to 100% utilized across all three products, whereas Dewas has some spare capacity for Chemical A.

Figure 15.7 Production Capacity Utilization—Baseline Versus Scenario with 20% Growth

When we look at the customer service aspect of this scenario (see Figure 15.8), this shows another key aspect that is important to consider when evaluating new alternatives.

 

Figure 15.8 Services Levels (% of Demand Served) by Distance Band

This shows that service levels were similar for the 800km distance bands, but started deteriorating for longer distances. This makes sense as the model would try to serve demand within short distances first, and serve remaining demand from wherever capacity is available.

The preceding scenario tells us and the management team a few things:

  • ■ There is capacity available within the current network to support 20% demand growth.
  • ■ However, this will mean higher logistics costs in order to utilize available capacity. This would also have a negative impact on customer service levels, especially for customers located beyond 800km from plants.
  • ■ This also tells us that there is very limited spare capacity available to handle demand spikes higher than 20%.

Given the capacity sensitivity, the management team wanted to understand the best location(s) to open a new plant. To answer this question, we will add a list of potential plant locations based on customer and demand demographics. For the purpose of this analysis, we will add a list of 30 locations to the model. The list of locations and their placement on a map is shown in Figure 15.9.

Figure 15.9 List of Potential Plant Locations

For the new plant(s), there are three main questions we are trying to answer:

  • Where should the new plant be located?
  • What products should be manufactured at the new plant(s)?
  • How much capacity needs to be added at the new location(s)—that is, how many production lines do we need to add?

In terms of production capability, we want the model to determine what to make at the new plant(s), so we will give the potential plants the ability to manufacture all products.

As far as capacity is concerned, there are a couple of ways to model this:

  • We could give the new plants unlimited capacity to produce any products. This will allow the model to determine how much product should be produced—we can then determine how many lines are needed based on this.
  • We could add capacity representing a certain number of lines, and let the model utilize this available capacity.

Both options are valid and applicable to model capacity. The best approach will depend on the nature of the business, capital equipment costs, and so on.

The advantage of Option 1 is that it would give the model the most flexibility in determining product sourcing volume from the new plant. Alternatively, this could also be viewed as a disadvantage because the model may come up with a volume mix that represents 1.05 lines—that is, 5% capacity utilized for a second line. It will probably not make sense to open a second production line to handle a very small incremental spillover volume.

Given this, we will use Option 2 and model a specific number of lines at each potential plant and see how the model utilizes this capacity. To give the model sufficient flexibility, we will give each new plant three potential lines for Chemicals A and B, and four potential lines for Chemical C. This allows the new plants to be potentially much larger than the existing plants.

For now, we will not add any opening costs or fixed costs for the new lines—we will look at this a bit later. As we have seen in earlier chapters, we can evaluate the impact of fixed costs without having to model them directly in the model.

Figure 15.10 shows the existing and potential plants with the available lines and capacity.

Figure 15.10 Production Capacity by Plant for Current and Potential Plants

We will also need to add transportation rates and lanes from the new plants (and attached warehouses) to all the customers. We will also create a plant site grouping for all potential plants, allowing us to apply constraints on the number of new plants that can be opened.

We are now ready to run our new set of scenarios with the new plant locations. The first scenario we will run would be to allow the model to select up to one new plant location.

From the map shown in Figure 15.11, we can see that the model has opened a new plant in Dhanbad (Jharkhand) that is serving customers in the eastern part of the country, and the Lucknow plant is serving customers in the northern region only.

Figure 15.11 Map Showing Results of Scenario with One New Plant

When we look at the costs comparison (see Figure 15.12) against the scenario without new plants, we see that the overall costs reduced by 5%—the savings are attributed primarily in outbound transportation costs because the model now has sufficient capacity to serve demand regionally.

Figure 15.12 Cost Comparison for Scenarios With and Without New Plant

The production capacity analysis (see Figure 15.13) shows that at the new plant the model opened two production lines each for Chemicals A and B, and three lines for Chemical C. We can see that three of the seven lines opened are highly underutilized. This is because the model did not have any penalties to open a new line, so it opened a new line if it was cheaper to serve even only a handful of customers from the new plant. We know that it would not make sense to open a new line that is only 5% to 10% utilized. We could easily address this problem by adding a fixed opening cost for each line.

Figure 15.13 Comparison of Production Capacity Utilization Between Scenarios

We will add a fixed cost of Rs 5 million for each line based on the average cost for a new line.

The results of this scenario are shown in Figures 15.14 and 15.15. The chart in Figure 15.14 shows that only one line is opened in the new Dhanbad plant when we add a fixed line cost. The volume served by Dhanbad in the previous scenario is now picked up by Aurangabad and Dewas plants as we see their capacity utilization going up significantly. The consequence of this shift is seen in the cost comparison table. We see that the outbound transportation cost goes up by almost Rs 5 million when we add the fixed line cost at Dhanbad.

Figure 15.14 Production Capacity Utilization Comparison—New Plant With and Without Fixed Line Costs

Figure 15.15 Scenario Results Comparison—New Plant With and Without Fixed Line Costs

Note that comparing these two scenarios is not an “apples-to-apples” comparison because they do not have the same cost structure. However, we are doing this to help gain insight on the impact of adding a new cost component on the solution as previously developed. The management team can use this insight to understand the key drivers in their supply chain, and thereby make more informed decisions.

 

Single-Sourcing

The management team had received feedback from some of their customers that they would prefer to receive all their shipments from one plant only. As we can see in the baseline map, there are several customers that are served from more than one plant. This multisource customer assignment was also turning out to be a challenge operationally for the logistics department. So the management team was interested in understanding the impact of single-sourcing individual customer locations.

Let’s run a new scenario allowing an additional plant but with the constraint that customers must be single sourced.

The map in Figure 15.16 shows that the new plant in Dhanbad is still valid. The scenario costs (see Figure 15.17) show that there is a less than 1% difference in total costs due to the single-sourcing constraint. This shows that the new network is capable of handling customer single-sourcing with relatively no impact on logistics costs. This has a very positive impact on operational complexity because each customer can be assigned to a plant for order processing and shipping.

Figure 15.16 Map Showing Results of Scenario with Customers Single-Sourced

Figure 15.17 Results Comparison of Scenarios With and Without Customer Single-Sourcing

State-Based Single-Sourcing

To evaluate certain tax policies and their implications, the management team wanted to understand the impact of serving each individual state from one plant only—that is, single-sourcing of all customer demand within each state.

To do this, we have created customer groupings by state and applied a constraint in the model that forces each customer group (state) to be sourced from only one plant. The map in Figure 15.18 shows the results of the model with each state single-sourced from one plant/warehouse only. First of all, we notice that the new plant location is not Dhanbad anymore—it has moved east to Howrah instead. Also, we see that the customer assignments look a bit odd—the Madurai plant takes over serving some customers much further North than its previous territory, as well as the new Howrah plant serving customers all the way in the Northwest.

Figure 15.18 Map Showing Results of Scenario with States Single-Sourced

Why is this happening? As discussed in Chapter 5, “Adding Capacity to the Model,” we are creating a harder knapsack problem when we aggregate demand into larger chunks. In this case, the model is forced to look for a solution in which each plant supplies the entire demand of the states it serves. Because capacity is tight, we may get answers that look strange.

Even worse than a map that looks strange, we can see from Figure 15.19 that with this constraint JPMS can only meet 98% of the demand. The capacity constraints combined with the constraint to serve a state from one plant do not allow the flexibility to meet all demand. This analysis gives the management team at JPMS some idea of what kind of impact this constraint may have on their supply chain.

Figure 15.19 Cost Comparison of Scenario with States Single-Sourced

A quick look at the production capacities by plant by line (see Figure 15.20) shows a picture that is a bit hard to understand without reminding ourselves about the sourcing constraints. The scenario on top represents the previous solution with the new Dhanbad plant. Each plant is highly utilized relative to demand in its region, and one new line opened at Dhanbad. When we ran the scenario with each state being single-sourced, we see that some of the lines are fully utilized whereas others are hardly used. Why is this happening? First of all, the demand distribution across the states is not similar for all products—that is, some states have higher demand for Chemicals A and B and very little demand for C. The model is factoring this constraint, along with transportation costs, and fixed line costs to come up with the capacity utilization profile.

Figure 15.20 Production Capacity Comparison of Scenarios With and Without Sourcing Constraints

To reiterate our point from earlier, we see that this single-sourcing constraint added a high level of complexity to the model. At the same time, it also provides us with an opportunity to understand the dynamics of the supply chain and how different variables impact each other.

Based on the multiple scenarios run so far, as well as others run to test other potential plant locations, the management team had sufficient information needed to make a sound decision on their manufacturing strategy. Note that we have focused primarily on key supply chain costs that were considered part of the network model—there are several other qualitative factors that impact the choice of a new plant, such as tax benefits, proximity to key customers, availability of skilled labor, and so on.

The management team will need to consider these factors before they make a decision; however, they can combine these non-quantifiable factors with good cost and service data to understand the value of these the non-quantifiable factors.

Lessons Learned from the Case

We developed a model focused on manufacturing capacity and analyzing the best location for a new plant. We could see that capacity constraints were important to understand the key questions that we were trying to answer. However, the capacity constraints can also make the model far more complex, thereby yielding results that may not be intuitive or make sense. This is especially applicable when we start applying sourcing constraints at the aggregate level (for example, by single sourcing a state). In addition, we also reviewed the importance of modeling fixed line costs and how this impacts decisions regarding how many new lines are needed to support demand.

Grading Criteria

 

Criteria Non-performance Basic Proficient Distinguished
Evaluate a supply chain network’s ability to scale for changing market conditions.
20%
Does not summarize a supply chain network’s ability to scale for changing market conditions. Summarizes a supply chain network’s ability to scale for changing market conditions. Evaluates a supply chain network’s ability to scale for changing market conditions. Evaluates a supply chain network’s ability to scale for changing market conditions; provides effective strategies to scale operations.
Evaluate the financial implications of various scenarios in supply chain network designs.
20%
Does not summarize the financial implications of various scenarios in supply chain network designs. Summarizes the financial implications of various scenarios in supply chain network designs. Evaluates the financial implications of various scenarios in supply chain network designs. Evaluates the financial implications of various scenarios in supply chain network designs; recommends actions to improve financial results under different scenarios.
Evaluate the customer service implications of various scenarios in supply chain network design.
20%
Does not summarize the customer service implications of various scenarios in supply chain network design. Summarizes the customer service implications of various scenarios in supply chain network design. Evaluates the customer service implications of various scenarios in supply chain network design. Evaluates the customer service implications of various scenarios in supply chain network design; recommends actions to optimize service.
Evaluate the operational complexity implications of various scenarios in supply chain network design.
20%
Does not summarize the operational complexity implications of various scenarios in supply chain network design. Summarizes the operational complexity implications of various scenarios in supply chain network design. Evaluates the operational complexity implications of various scenarios in supply chain network design. Evaluates the operational complexity implications of various scenarios in supply chain network design; recommends scenarios that manage complexity while meeting other goals.
Write in a scholarly manner by providing validation and scholarly evidence through writing that is free from grammatical and mechanical errors and adheres to APA style.
20%
Does not write in a scholarly manner. Writes in a scholarly manner and provides validation and scholarly evidence, but writing, grammar, mechanics, APA style, and overall output needs improvement. Write in a scholarly manner by providing validation and scholarly evidence through writing that is free from grammatical and mechanical errors and adheres to APA style.Writes in a scholarly manner by providing validation and scholarly evidence through writing that is free from grammatical and mechanical errors and adheres to APA style. Writes in a scholarly manner by providing validation and scholarly evidence. Writing is free from grammatical and mechanical errors, adheres to APA style, is concise, and presents ideas cohesively and logically.

 

ly evidence through writing that is free from grammatical and mechanical errors and adheres to APA style.
20%

Does not write in a scholarly manner. Writes in a scholarly manner and provides validation and scholarly evidence, but writing, grammar, mechanics, APA style, and overall output needs improvement. Write in a scholarly manner by providing validation and scholarly evidence through writing that is free from grammatical and mechanical errors and adheres to APA style.Writes in a scholarly manner by providing validation and scholarly evidence through writing that is free from grammatical and mechanical errors and adheres to APA style. Writes in a scholarly manner by providing validation and scholarly evidence. Writing is free from grammatical and mechanical errors, adheres to APA style, is concise, and presents ideas cohesively and logically.

 

Please follow and like us:
error