Logistics Optimization
Logistics is the bloodline of any Supply Chain
Logistics is the bloodline of any Supply Chain and it became an important part of every growing economy. Indian Logistics sector is growing at a very high pace with support from market forces like e-commerce, 3PLs and government forces like GST, expressways, PM Gati Shakti etc. Every organization wants to reach the customer faster so that on time delivery targets are met and logistics plays a major role in fulfilling it.
Distribution is one of the 5 core processes of supply chain, the other four being Procurement, Production, Storage and Order/Demand Management. Distribution is a very critical process as it is closer to the consumption point and maximum value has already been added to the product in all its preceding processes.
Considering an organization with standard network of distributors and retailers, following are the different distribution types that are largely followed at various stages of distribution in order to move Goods from Plant to the Customer.
- Point to Point Distribution (Ex: Plant to RDC, RDC to CFA, Plant to CFA)
- Multi Drop Distribution (Ex: DC to Retailers, Plant to CFAs/Stockists)
3. Last Mile Distribution (Ex. Retailers/CFAs to Customers)
Let’s focus on the Point to Point Distribution operations that most of the organizations try to improve by planning and executing them in an efficient manner since the stakeholders are part of the same organization. Dispatch/Logistics Planning team is responsible for planning the distribution of goods with a single line objective of fulfilling the requirement as per orders/forecast at stock points (RDCs/CFAs).
As we know, Planning is all about using resources effectively while trying to meet the objective of the plan, planning teams try to achieve their objective by while improving the performance metrics that are key to the Distribution. Typical KPIs that are looked at are
- OTIF — How much requirement is met On Time In Full.
- Availability — Target % to be met by a particular date as per monthly forecast
- No. of Full Truck Loads — Minimize the number of trucks required
- Truck Utilization — Maximize the truck utilization
- Cost/Unit — Minimize the distribution/fulfillment cost
- % of Direct Dispatches — Improve the direct dispatches to demand point to avoid transshipment, material handling and storage at intermediate warehouses.
A common scenario where distribution planning is required is in case of finished goods movement over the network of plants and Hubs/RDCs/Warehouses/CFAs. All these movements will be of high volume for all the products. Distribution should be planned efficiently such that right product is available at the right location at right time.
Hubs/RDCs are bigger in size and handles large volume of products and typically cover an entire region like South India whereas CFA/Warehouses are smaller in size to handle a particular city/state requirement. Typically, a plant manufactures a product in bulk and does not produce all SKUs at the same time. And a CFA will not have huge requirement of single product since it caters one small area, but has lesser requirement for more number of products. That’s why an intermediate location called RDC is used where products from multiple plants will come and a full truck is formed at this location with all the products and send it to a CFA. Arriving at the Dispatch plan manually using spreadsheet is the common scenario in all organizations and it is getting complex because of huge product variants, multiple channels, product-region level promotions etc. As the complexity goes up, manual plans become inefficient and time consuming.
Planners use rule of thumb while trying to generate truck loads by using filler materials so that they can make it full and dispatch the actual requirement as early as possible. This increases the inventory at the end location with unnecessary product. It also increases inter CFA transfers to balance the stocks which are excess in one location. Manually determining full truck loads for say 30 locations and 100–300 products will be challenging and time consuming.
Decision making technologies are useful in handling all the complexity and generating an optimal/best Dispatch plan given the various limitations/constraints. You might ask, what complexity we are talking about apart from the data volume. To give a glimpse, look at the different decisions that it can make as listed below in the order of increasing complexity
Decision#1: Calculate the Net Requirement for each product and location combination by considering the stock on hand, goods in transit, open orders and netting it with forecast and safety stock requirement across multiple days into future horizon
Decision#2 — Selecting the cost-effective source by exploring the entire network graph to fulfill the net requirement of each product-location combination by checking the lead time, truck availability and material feasibility
Decision#3 — Dispatch the right product in right quantity using right mode to right location to maintain service levels, inventory norms and minimize stock-outs and finally to minimize the total distribution cost.
Decision#4 — Prioritize the dispatches to maintain service levels of various channels and priority customers.
Decision#5 — Determine the optimal load and frequency of each truck type to each location while maximizing the truck utilization and adhering to weight and/or volume limitations.
Decision#6— Stock distribution to demand locations in a fair share mode incase of limited material availability to minimize sales loss.
I just listed general requirements which many organizations look for in a advanced planning solution and they will have their own set of constraints specific to their business and operations which are managed by approximations in excel.
It still might look like a theoretical challenge that you read in many case studies or papers. And so, in the next post let me present you a real distribution problem and how it was solved and implemented using optimization techniques.
