The Digital Agenda: Autonomous Systems

by | Iter Insights

Autonomous Systems

autonomous systems iter insights featured image

In the second of our four “digital agenda” articles we look at Autonomous Systems and their associated benefits and risks.

Autonomy is increasingly talked about in the context of future systems and their operation, so it makes sense to understand a little more about them. As a start point let’s define what we mean by autonomous systems in the context of supply chains …

“An autonomous system (AS) is a network or a collection of networks that are all managed and supervised by a single entity or organisation. An AS is a heterogeneous network typically governed by a large enterprise” (Wikipedia).

The key phrase here is “connected and managed by a single entity”. But this definition didn’t fully meet our expectations, so we searched again and found: “An autonomous robot is a robot that performs behaviours or tasks with a high degree of autonomy (without significant external influence)”. Wikipedia again.

Taken together these definitions get much closer to our view of what an autonomous supply chain is: connected, governed by a single entity and operating with a high degree of autonomy.

So, what does this mean in terms of modern global supply chains?

Hyper-linking, hyper-competition, and hyper-turbulence are typical phenomena of “real-time economies” in a world of diversity and change (Tapscott, 1999; Siegele, 2002).

Supply chains need to be increasingly complex but also increasingly flexible to provide levers that can be used to respond to events and trends. This leverage is powerful but requires great knowledge and experience to ensure it delivers benefit rather than noise.

Let’s consider some key drivers for increasing use of autonomous supply chain systems:

Increasing Complexity

Whilst it is important that we don’t make our supply chains more complex than necessary, we are driven to increasing complexity by our environment, and greater segmentation. This drives the need for autonomy to make supply chain systems manageable and to keep their operating costs down. This autonomy needs to be clearly underpinned by clarity at a fundamental level as to what you need your systems to do.

Speed of Response

Autonomous systems can sense changes and react far quicker than humans. Good examples are emergency braking systems in cars that can sense the vehicle in front has changed its speed a second or two before a driver can. A more extreme example is the Eurofighter Typhoon (seen above) which has a foreplane delta design which is inherently and intentionally aerodynamically unstable in subsonic flight. This provides enhanced manoeuvrability but requires a complex flight control system to support the pilot as the computer systems can react more quickly at lower speeds. In a similar vein, complex systems may become unstable if you cannot react rapidly enough, again driving the need for autonomy.

Always on Supply Chains

Global supply chains need to respond 24/7, 365 days per year and as they do, planning cycles get shorter and shorter. This is demanding for planning teams to keep pace with and therefore they benefit from any support that allows them to focus on the “important few rather than the many”.

Cost to Serve

Cost to serve modelling/optimisation has traditionally been carried out periodically and decisions made as a result. Autonomous systems can make these decisions in almost real time, providing improved customer service at a reduced cost provided flexibility is built into supply chain design. This is increasingly important as channels multiply and the environmental and sustainability need to be included as “costs”.

Self-Healing Supply Chains

Self-Healing Supply Chains bridge the gap between supply chain planning and execution and spot significant issues and take corrective action within agreed tolerances before they impact performance. They can also provide corrections for early design assumptions over time and significantly reduce supply chain Time to Recover from major disruption.

Customer Insight

As sales operations increase in sophistication companies gain increased insight into customers taste and buying habits, allowing addition items to be suggested when ordering. Managing the demand swings that result requires sophistication to ensure the real drivers of that demand are understood and managed correctly.

We can understand how AS add value but there are major changes required in respect to planning strategies, operating policies and rules and most significantly in respect of data and the culture.

This can be summed up in the phrase “Right not Almost Right” and if we are going to let our systems behave autonomously, we need to ensure the data within them is accurate and the rules and tolerances fully define how they will operate. This needs careful consideration now not in future as it requires significant time and effort to get this right. It’s not enough for your self-driving car to turn near the corner, it needs to turn at the corner!

This is challenging in existing systems where you easily drill down into transactional data planning policies and understand why actions were taken and decisions made but it is almost impossible in the world of AI and deep learning which become increasingly black box solutions. We will consider this along with other technologies in the next article.

If you are struggling to come to terms with what this may mean for your organisation or have already started on the journey to implementing autonomous supply chains and need an independent perspective to balance the input of your teams/systems vendor please contact us by clicking here.

Colin Prout
Delivery Director
Iter Consulting