Firstly, what exactly is the Physical Internet?
The physical internet, abbreviated frequently to PI, allows for optimisation of transport and logistics through connecting the fixed facilities (such as warehouses) and dynamic infrastructure (such as pallets) to the digital world, so that transport of goods can be ‘self-organised’ as transport flows are able to respond to real-time demand and supply. A PI system is also, via real-time tracking, able to respond to real-world transport constraints and delays.
I can see the benefit in knowing where freight is, if it is running late and if so, doing something about that, but beyond that, what would be the justification in terms of time, energy, and cost?
There are many benefits, some immediate and others longer term. For an individual company, the detailed logistics of their complex supply chains can be managed through a third-party system, removing the need to invest in new solutions to track, trace and divert flows of goods. Tracking also allows for synchronisation of manufacturing, warehousing, port and retailer or other service user activity, improving workflow along the supply chain. However, the real economies are realised through the opportunity to collaborate and consolidate goods, so reducing empty running, CO2 and other environmental impacts, improving asset utilisation and improving the cost of transport for all users. Such an integrated approach also increases the viability of multi-modal logistics chains, better combining the existing networks of rail, road, sea, and inland waterway.
Surely 3rd party logistics companies are well placed to consolidate and find economies?
Logistics and transport providers do their best to consolidate and combine flows, most obviously in areas such as shipping and rail freight booking. And whilst more broadly, most LSPs look for synergies such as backloading and other two-way flows, they are often focused on the operational performance of each individual client’s supply chain as a separate, sub-optimised solution. The closest today to a shared network would be the pallet networks, and perhaps postal systems, both of which offer a many-to-many standardised solution.
Ok, sounds good – so how does PI logistics work?
PI (physical internet) logistics optimises collaboration through replicating the physical world as a digital simulation. The real time simulation gathers data on demand from many users. For example information on transport orders, items, origin, destination, handling requirements (such as weight, dimensions, stacking constraints, temperature control, rules for hazardous goods, etc), lead time and other service level requirements, plus any customs process details.
The simulation also gathers information on capacities, for example, information on available road, rail, shipping, port, or warehousing or cross-docking facilities, including detail on vehicle fill, container fill, rail wagon, and warehouse fill rates. To enable the matching of users with operators, every demand and supply in the network has a digital twin.
Hold on a minute, what is a digital twin?
A digital twin is a digital copy of a real-world demand (such as pallet of goods that needs to be moved) and supply (such as available space in a container that is being moved by rail and road from origin and destination).
Each item moving through the network is replicated as a unique digital identity, and the digital twin becomes attached to an automatically generated and unique audit trail (often called blockchain). The whole network can then be simulated in real time, as a digital twin of the real-world infrastructure and network.
How is data collected and integrated?
Data is derived from existing legacy data flows, via API (integration software or ‘middleware’) that enables existing systems to talk to each other. Data is also derived from monitoring real world actions, through sensors and connectivity built into the real-world transport infrastructure, built into containers, (including, for example, chilled goods stillages), warehouse or port facilities, and collected via the existing connectivity of ships, trucks, and trains.
Some of the data required by the simulation is static – such as the total usable capacity of a warehouse – and some is dynamic – such as transport delays and congestion, or the availability of empty space within a warehouse or scheduled train route. The simulation runs algorithms that optimise the overall use of the system, so that the consolidated flow of goods drives a ‘self-organisation’ of the transport planning.
How does the system ‘self-organise’?
Cloud-based virtual ‘hubs’ create marketplaces that gather and calculate the best overall organisation of resources. Standardised protocols (much like TCP for the internet), help any point of origin ‘hub’ to exchange information and integrate simulations with others along the delivery chain towards each transport demand destination, to create a potentially unlimited and scalable network of networked solutions. The algorithms are deployed as on-demand ‘PI services’ that effectively work as a series of problem-solving processes, that together culminate in the creation of an executable transport plan that also, via the data attached to each transport order, becomes an audit trail of the logistics process.
What types of real-world problems can networked PI applications help solve and support?
The ICONET project has built and tested the application of PI technologies and solutions in real world environments, all of which can be combined as components within a PI network. ICONET is a Horizon 2020 EU-funded project that involves manufacturers, retailers, IT and logistics companies, tasked with proving the PI concept. Four project tests, or living labs, replicated real world logistics problems, and have successfully shown the benefits that PI can bring to commercial operations.
The first lab was a port, and the PI solution involved optimising operations internally within the port, so making better use of the rail system within the port authority area. However, the PI solution also enabled synchronisation of process tact-times and flows of inbound and outbound traffic, both to sea, and deep within the hinterland of the port. A port is a natural place to try to synchronise collaborative activity, as the multiple users, competing freight forwarders and carriers, must overcome significant barriers to trust resulting from deeply ingrained conflicts over prioritisation and access, partly an outcome of common bottlenecks and delays.
The second lab involved collaborative use of a high-volume corridor, a trucking route, that generated cost savings in moving consolidated freight over long distances at greater speeds and frequencies. The automated consolidation and sortation (so optimised loading and packing) also produced greater load efficiencies, alongside cuts in fuel consumption, driving hours and emissions.
The third lab demonstrated shared use of a distribution warehousing network, for managing safety buffer, tactical replenishment, and seasonal or cyclical stocks across a market area. This solution allowed storage space to be rented as required by multiple users at multiple locations, optimising resources for warehouse owners, operators, and users.
The fourth lab arrived at a timely point, enabling a retailer to optimise last mile logistics during the jump in demand for home shopping resulting from the Covid-19 crisis.
When serving the end customer, it is a strategic decision as to whether to share fast and efficient delivery services, or whether to operate a dedicated service offering brand differentiation; however, whilst a large retailer benefits from high volumes and drop densities, making shared flows less attractive, longer-term, once operations are established, they too would benefit from the backloading opportunities unlocked by operating as a multi-user network. All four labs effectively demonstrated the components and cogs that are required to build a wider and unlimited network of PI enabled shared logistics.
How to make this happen?
Collaboration is often viewed as ‘career limiting’ by many supply chain managers for good reason, as it tends to expose weaknesses and bottlenecks, and inbuilt underutilisation does provide a degree of insurance against failure to meet contractual targets. However, the physical internet is designed to be capable of quickly finding additional resources and capacities to avoid delays, breakdowns, or loss of customer service, and to make a collaborative solution as low risk, secure and easy as possible.
A PI system provides greater accuracy for delivery times and allows mitigating action to be taken in real-time, often before a delay is experienced, through rerouting of flows. Getting companies to join such a network will also require such guarantees, and so is likely to be delivered via low cost, low commitment subscription, and easy integration through open API. Like all great networks, the growth of users and traffic itself generates better outcomes for all participants, but early adopters will have the opportunity to learn how to use it most effectively.
Written by Ben Waller and Stephen Rinsler ©Ben Waller, Stephen Rinsler. This article was first published in Focus, the house journal of the Chartered Institute of Logistics and Transport (UK) Ltd.
ELUPEG Project Team part of EU Horizon 2020 Grant Funded Project: ICONET Consortium