SMART OPTIMISATION TECHNOLOGIES

1. DESCRIPTION OF TECHNOLOGY AREA

The application of heuristics and metaheuristics to resolve optimisation problems requires the statistical study of the results. Statistics are also used in the parameterisation of algorithms.

When there are algorithms with parameterisation values, finding the best combination of values of those parameters to solve a problem is, in turn, an optimisation problem. The process of finding the best parameters for an algorithm is called calibration.

Usually, the calibration process is long and expensive (in terms of calculation), since it requires generating large amounts of results for further statistical study.

This statistical study allows us to determine the best parameters for an algorithm, so that it can solve a set of problems with certain characteristics. This also means that, if the characteristics of the problems under study change, the parameterisation of the algorithms should be changed. That is, the calibration of the algorithm should be performed every time a change in the input data of the problem under study is detected.

Thanks to the technological advances in the area of computing, doors have been opened to develop the automatic calibration of algorithms. This requires a closed system capable of performing the calibration study of an algorithm quickly and automatically without the intervention of a person.

Big Data and cloud computing technologies allow the development of this type of automatic processes, which also require the use of Artificial Intelligence.

2. EVOLUTION

Today and in the near future, the need to manage large amounts of data, remote access and distributed computing are encouraging the development of cloud optimisation technologies.

A distributed optimisation system can allow parallel computing and will also allow access to information, results and indicators from all points that use the resulting information (even from other optimisation systems).

The availability of cloud optimisation systems will scale optimisation to levels that allow to integrate material suppliers, manufacturers, distributors and final product vendors into a single connected system. This integration, in turn, will allow to develop the global optimisation of shared resources throughout the distribution chain.

3. USE CASES AND APPLICATIONS

Some areas of application of combinatorial optimisation are:

SCHEDULING PROBLEMS

Scheduling problems are a large family of problems applicable to almost any industrial organisation and even to project management.

• Job-shop, flow-shop, parallel machines, hybrid flow-shop
  Various problems dealing with the optimisation of manufacturing tasks in different types of industrial plants.
• Project scheduling

This set of problems tries to find the best order of execution of tasks in a project and is closely related to other types of scheduling. Many algorithms proposed for flow-shop scheduling (or its variants) can be applied to project scheduling as well.

TSP

Travelling salesman problem, which consists of finding the best route (shorter, faster, cheaper, etc.) that passes through a set of points (cities) only once.

VRP

Vehicle route problem. It is similar to the TSP, but takes into account that the route to be generated must be a closed one, that is, it must end at the same starting point.

There are many variants of TSP and VRP in which you can specify several vehicles, heterogeneous vehicle fleets, delivery windows, product deliveries and collections, etc.

Organisational problems

• YCSP: Yard Crane Scheduling Problem. This problem tries to find the order of treatment of containers that enter and leave the yard, such that the final time, the route of the crane or the movements of containers, etc. are reduced.
• Marshalling and pre-marshalling: Problems that involve accommodating or rearranging containers in a port’s container yard.
• Quay gantry scheduling and Ship or Vessel scheduling: Problem related to scheduling the cranes that load and unload the ships
• Berth allocation problem: Problem related to space allocation to the ships in the port