ModelEnterprise - concepts & technologies

How does ModelEnterprise work?

ModelEnterprise uses mathematical optimisaton techniques to solve a mixed-integer linear programming (MILP) problem.

The underlying model being solved by the MILP problem is defined as a Resource-Task Network (RTN), a representation that was coined at Imperial College London (Pantelides, 1994) and is now used for most such problems.

ModelEnterprise uses a common data model, meaning that a single enterprise representation can be used to model the entire enterprise from multiple sites to individual production units.

ModelEnterprise is NOT a rule-based planning system. Its capabilities go far beyond the limitations of such systems.

Advantages of the ModelEnterprise approach

The key advantages of ModelEnterprise are:

it provides rigorous mathematical optimisation, meaning that the optimal solution found is indeed the best one that can be achieved using the model.
a constraint can never be violated and thus the end operational plan will always be feasible
the same model can be used for many different cases and analyses, meaning that any initial investment in modelling can be spread over a number of applications.

Concepts explained

Some of the key underlying concepts that differentiate ModelEnterprise from other software are described below.

Optimisation

Complex production schedule for lubricants plant

A one-year production schedule for a complex lubricants manufacturing operation, showing optimal asset utilisation. Click on image to enlarge

A simple Resource-Task Network

(below) A more complex Resource-Task Network (courtesy Nexia Solutions)

The synthesis of the enterprise model results in a mathematical optimisation problem with an economic objective function.

The optimisation is usually aimed at maximising profit or minimising total cost of operations while satisfying a single or multiple targets (constraints) by manipulating, for example, the timing of operations or the amounts produced, subject to any penalties that might apply for late delivery.

Economic objectives are usually achieved by increasing asset utilisation and throughput to maximise capacity, and prioritising production according to economic criteria.

At no time will the optimisation ever violate the constraint on the available resources, such as material and operator availability, warehouse capacities, and equipment capacities, availabilities and usage restrictions.

The Resource-Task Network (RTN) concept

An RTN is a set of transformations that involve a resource (for example, "Dirty flask") being transformed by a task (for example, "Flask cleaning") to provide a new resource ("Clean flask"). This new resource can be the input to the next element in the RTN.

If the full picture were shown we would see another resource "clean water" transformed by the same task to "dirty water". A more complex RTN is shown below.

The RTN concept is a simple yet immensely powerful way of describing a complex system of resources and tasks. RTN elements are not limited to material transformation.

A "fresh operator", for example, can be transformed by an eight-hour shift into a "tired operator". "Raw material A in Basle" can be transformed by a transport "task" – for example, a train journey – into "Raw material A in Zurich".

All of the relevant equipment, recipe and demand information in a ModelEnterprise model is converted into Resource-Task desciptions automatically.

This creates a well-defined mathematical problem definition in a form that is easily solved by standard optimisation techniques. More complex Resource-Task Network

Enterprises and sites

ModelEnterprise typical enterprise – click to enlarge

Typical enterprise: a Europe-wide enterprise with 3 sites supplying 16 warehouses. Click to expand

ModelEnterprise Modelller – click to enlarge

An enterprise is a business entity comprising a number of sites at which materials are processed or stored. The enterprise also has contacts with external organisations from which it procures raw materials or to which it supplies finished products.

For example, the figure on the right shows a manufacturing enterprise that has three plants located in UK, Spain and Italy.

These sites supply a range of products to 16 warehouses distributed across Western Europe from which customer demand is met.

Using ModelEnterprise you can build detailed production models that realistically represent the production capacity at each site and the interaction between sites. This gives you the ability to make optimal decisions at the site and enterprise level.

Representation of resources

ModelEnterprise groups the production resources that are available at each site into three categories:

materials such as feed stocks, intermediates or finished products, or even emissions such as carbon dioxide and nitrogen oxides
equipment units such as reactors and mixers for processing materials; and tanks and silos for storing materials
common resources – limited resources other than materials or equipment that are used at that site such as manpower, steam or electricity.

A processing site also has recipes that describe how the production resources (materials, equipment units and common resources) are utilised by the production activities performed at that site.

The graphic on the right shows how the information is arranged in the ModelEnterprise Modeller project window.