A Non Mathematician's Guide to Optimisation
Maybe you are thinking about using optimisation software in your business.
Or perhaps you already are.
Did you know that different methods of optimisation exist, with different objectives and hence outcomes?
Artificial intelligence, neural networks, genetic algorithms and mathematical optimisation, while these terms sound like they belong in a science fiction movie, in reality they are some of the different optimisation methodologies (‘the engines’) used to provide different ‘optimised’ solutions.
So for those of us without a background in mathematics, computer science or operational research what does this mean?
Imagine that you have to construct a plan for your companies operations. However, there are many viable possible plans (those which fit the companies capacities and business rules). Your goal is to find the best plan for the company from all these potential plans.
Now imagine all of the possible plans as a landscape, with valleys, hills, and mountains of various sizes, where the ‘higher’ the landmark the better that plan is for your company.
In this analogy, the highest mountain of all, ‘The Mount Everest’ represents the best/most valuable/most profitable plan.
To find the highest mountain (the best plan) you need to search all the possibilities.You could try to use past experience to find the best solution, however this is unlikely to work in a complex situation. In our analogy it might only represent a nearby hill. It is still a viable plan, but it's not the 'Mt Everest.'
Typically, there are too many possibilities to consider, for experience alone, to work and you will need help to achieve this task.
Artificial Intelligence and Optimisation
A more methodical approach using automation is the way to find better plans. For example, you might decide to continue your search using powerful artificial intelligence (AI) optimisation tools (software). This software may use algorithms such as:
- Neural Networks (a computer model that tries to imitate the behaviour of connections between neurons in a human brain),
- Genetic Algorithms (a computer model based on the theory of natural selection and evolutionary biology).
Unfortunately, though these AI algorithms will usually produce an improved solution, most times they will only find a higher place (better plan), not the best plan. On rare occasions, they might locate ‘Mt Everest’, but most of the time, with these forms of optimisation AI, there is no guarantee of finding the best/highest solution.
Mathematical Optimisation is different!
Mathematical optimisation, where it is possible to use it, is always guaranteed to find the best possible solution from a set of available alternatives, in this case, the highest mountain, the ‘Mt Everest’. For example, it is possible to apply mathematical optimisation to the task of finding the best plan for aquaculture businesses.
Mathematical optimisation algorithms exploit the mathematical structure of the problem to filter the plans. So in this analogy, mathematical optimisation would filter the set of mountains to be only those over a certain height, and this makes it practical to always find the 'highest' landmark, the 'Mt Everest.'
What does this mean for my company?
If you use mathematical optimisation, then you will always find the best plan for your company, and this not only improves the companies results, but it removes much uncertainty in decision making.
Mathematical Optimisation Software for Aquaculture
For aquaculture businesses of all sizes the number of decision choices rapidly escalate, and it is simply too time-consuming for an individual to evaluate or to even see all the possibilities. Neptune not only finds the best solution, but it enables you to interactively explore other scenarios in seconds.
Dr David Wright, CEO PanLogica
Neptune is the most powerful planning software for the Aquaculture industry because it uses mathematical optimisation to always find the best plan for the entire value chain of the producer in a single step.
PanLogica was founded in 2001 when Dr David Wright identified the decision making complexities associated with the lengthy process of fish farming and created a software program, Neptune, that would make aquaculture business decision making more systematic, robust and time efficient.
We invite you to contact us for a more detailed explanation of Neptune and PanLogica.