subject: Shed Ahead [print this page] There are many real world applications for optimisation technology within supply chains given their complexity, scale and dynamics. Over the last 15 years, applications such as supply chain network optimisation (which optimise supply chain structures, resources and flows), inventory optimisation (which optimise inventory strategies based on service level, cost and risk trade-offs) and transportation route optimisation (which optimise transportation schedules based on service requirements, capacity and cost) have become more commonplace in market leading organisations.
These tools are becoming necessities as service, cost and asset efficiency competition increases, and advancements in technology have made them more accessible.
But in all of this, where does the warehouse or distribution centre fit? Sure, supply chain network optimisation techniques might help with the size or location of the facility. Inventory optimisation might determine the optimal amount of inventory to hold and replenish. And transportation route optimisation might determine which vehicle turns up to collect a load, but how do we optimise the layout of a facility and the processes within the facility?
Until recently, there hasn't been a good answer to this question. Optimisation can be found within systems that help to work out where and how big the four walls' should be, but of the systems and processes that work within the four walls, there has been scant trance.
There a several reasons for this. One reason is that warehouse operations aren't often seen as a high value' proposition. Many hold that (once the facility investment has been made and becomes operational) labour cost savings are the only real opportunity within owned (versus 3PL) facilities.
Another reason is that warehouse operations are "execution only" and only require transactional systems to ensure everything is in its right place.
A final reason (for this article, anyway) is that, if you get right into the detail of all the real world elements that come into play, there are literally millions of combinations of options to be considered. So it gets placed in the too hard' basket.
For example, when designing the layout and/or configuration of a warehouse, there are many factors to consider. Ideally, you're after the lowest cost (initial and on-going), highest capacity (storage and processing), most efficient and most responsive design possible. To inform this decision, one needs to consider demand hit' frequency and quantities by product, product dimensions, storage media and configuration options, facility dimensions, pick strategies, unique constraints and costs associated with products (eg refrigerated or special handling requirements), travel distances and time, labour capacity and pick efficiency, process stages such as packaging or order consolidation, potential bottlenecks, space allocation to receipt, despatch, staging, replenishment, bulk and storage/picking, safety considerations and regulations, order profiles (eg are some products often ordered together?), service requirements and product range/business growth considerations.
If you take the above into account and then consider a business with 2,500+ products with different demand profiles, dimensions and constraints, the range of possibilities is enormous. As a result, the tendency is to use simpler and non-standard approaches that don't consider all the options. For example, it may not be possible to exhaustively analyse all layout or configuration options based on various order profiles using a spreadsheet. As such, a solution might be generated in a spreadsheet that determines the best location for each item by minimising cost and/or travel time for each individual item. In this case, the high frequency items are generally located closest to the staging/despatch areas (ie the forward' locations). However, if the high frequency items are rarely ordered together, and are often ordered with other lower frequency items, the layout/configuration that appeared to be ideal may in fact significantly increase the overall pick time and reduce service levels and efficiency. In this case, pickers might have to walk the entire length of the facility on a large proportion of the orders.
If we then move to the processes that occur within a warehouse, we might focus on dynamic slotting and daily pick strategies.
Slotting is the process of re-assigning locations within a warehouse on a more frequent basis. As most warehouse layout/configuration analyses are done at a point in time, we recognise that we may need to re-slot the warehouse periodically to respond to changes and maintain or improve efficiency. A simple example of this is seasonal products. If clothing ranges are sold on a seasonal basis, it may make sense to re-slot the warehouse when moving from the winter to the summer season. In summer, the warehouse will run more efficiently if swimwear is in the forward locations instead of the wool coats. An obvious example, but hopefully it illustrates the point.
Regarding pick strategies, the aim is to assign picking schedules that minimise costs and travel time.
At present, slotting and pick strategies are dealt with by Warehouse Management Systems (WMS). Transaction systems and WMS are fundamental to the performance of warehouses that exceed a certain scale/complexity dimension. WMS is critical to maximising storage capacity via variable location assignment, maintaining data integrity and supporting or enabling the detailed daily execution processes within facilities. However, the slotting functionality within most WMS offerings is relatively basic, and they typically do not have the hardcore mathematical capabilities to optimise daily operations.
And recently, we came across a German company called Axxom in our travels. Axxom provides an integrated suite of supply chain optimisation tools called ORion-PI including, supply chain network optimisation, advanced planning & scheduling (APS), facility optimisation and workforce optimisation.
As we looked into ORion-PI, we came to realise that it offered optimisation functionality for warehouse design and daily operations, and it became apparent that this was a new frontier for warehouse operations performance improvement in Australia.
From a warehouse design and configuration standpoint, ORion-PI analyses all the variables mentioned earlier (literally, millions of possible combinations) to determine the optimal facility layout from a cost, capacity and service perspective. Using the previous example, the layout would have been influenced by the products that are ordered together, and a more effective design would have resulted.
From a warehouse operations perspective, ORion-PI can optimise in several areas and at multiple levels of the process. For example, ORion-PI can optimise the sequence that orders are released into a facility to maximise throughput efficiency. It can optimally and automatically re-slot the warehouse as warranted by changing conditions. It optimises pick strategies and processes based on the order requirements over a specified time period (and this may change on a daily basis). It can optimise shipments by minimising the number of containers/ boxes used, filling containers/ boxes up to maximum volume or weight allowed and choosing the optimal size from possible sizes of transport units. In this case, ORion-PI automatically considers hazardous freight, fragile items, special freight etc and evenly distributes volume between transport units of one order. There is also replenishment optimisation which calculates the optimal position of pallets in reserve racks depending on the products' position in the shipping line considering replenishment frequency. This is simply a sampling plate' of what's available versus an exhaustive list.
These toolsets are not as data hungry' as many believe them to be, and they bolt on to existing ERP or WMS systems in the same way that "best-of-breed" Demand Planning systems do. However, Axxom AG has a direct and strong relationship with SAP, and there are existing integration tools between the systems, in addition to direct link' data import tools built in Excel for other systems.
If we now move back to the potential benefits and opportunities, these tools have proven themselves capable of significantly improving warehouse efficiency (eg cost per pick and number of picks per hour) and capacity. In addition, if we consider that growth sometimes necessitates additional facility investment, ORion-PI can reduce the capital intensity of growth, improve Return on Assets (ROA) and enable the product range to expand, as appropriate. Basically, it's way of your business avoiding additional expenditure where others may need to outlay capital to achieve the same outcome, whilst at the same time improving warehouse operations efficiency.
There are many organisations around the world using ORion-PI to optimise various aspects of their warehouse and distribution centre operations. Some of the more recognised global brands include DHL, Adidas, Avon Cosmetics and BMW.
However, there are very few examples of local organisation benefitting from these proven techniques. As such, there appears to be a step change' opportunity in warehouse performance available to Australian organisations.
