Skip to content

Objectives

An objective is what the plan tries to improve. Each problem chooses one or more.

With several objectives, the engine looks for trade-off solutions: plans that cannot be improved on one objective without worsening another (the Pareto front). Objectives can also be ranked by priority, so that an objective only breaks ties when the previous ones are equal, within a tolerance.

Each objective has a weight and, in some cases, its own parameters.

Finishes everything as early as possible. Minimises the moment the last task finishes.

When to use it: when what matters is freeing up capacity early and there are no committed dates.

Each order finishes as close as possible to its due date.

Parameter What it is
Lateness weight How much each time unit late is penalised
Earliness weight How much each time unit before the date is penalised
Duration weight How much it is penalised for the order to stay open for a long time
Weights per order Each order or customer can have its own

Requires orders with a due date. When to use it: make-to-order production.

Minimises the time between the first and the last task of each order: less work in progress.

Spreads maintenance stoppages so that not too many assets are down at the same time. Penalises the daily variation in the number of assets under maintenance and, above all, the peaks.

When to use it: fleet maintenance.

Each service has an allowed interval since the previous one of the same type (a minimum and a maximum). This objective places each service near the centre of that interval: not so early that interval is wasted, nor so late that it becomes a risk.

Changes as little as possible relative to a reference plan, usually the current one. Changes close to the start weigh more than distant ones: moving something from tomorrow is more disruptive than moving something from next month.

When to use it: when replanning.

Some resources have a set-up cost that depends on which task follows which (a change of colour, format or tooling). This objective minimises the sum of those costs. It is described with a table: for each resource, the cost of doing task B straight after task A.