Mining Goes Digital – Mueller et al. (Eds)
© 2019 Taylor & Francis Group, London, ISBN 978-0-367-33604-2
386
Understanding plan’s priorities: Short term scheduling optimization
A.B. Andrade
AngloGold Ashanti, Sabará, MG, Brazil
P.C.B. Rampazzo
Unicamp, Limeira, SP, Brazil
ABSTRACT: Three levels are usually applied to mining planning: Long-term sets the
strategy while medium and short-terms aim to achieve main targets (ore production, grades,
meters developed, metal production, etc.). Manually scheduling a monthly plan, which is a
widely used method, can be a difficult task: A 30-day plan can take days to be elaborated and
appropriately validated. In a dynamic mine environment where day-to-day compliance varies, after the first planned day, some of the tasks will not be executed, some might be delayed
and some anticipated. It will be needed to review the plan and so next day operation can be
correctly prioritized. Most of the time, rescheduling the full plan might not be an option and
suboptimal decisions are made. This paper proposes an algorithm for short-term scheduling
in underground mines with a flow shop scheduling with parallel machines model, optimizing
and automatizing short term planning process.
Keywords: mixed linear programming, short-term scheduling optimization
1 INTRODUCTION
Mining industry, as any capital intensive production system, requires detailed planning to
guarantee that resources will be allocated the best way and that returns on investments will
happen with predictability.
Usually mine planning is carried out in three different levels. Long-term will set the strategic main targets; medium and short-term, the tactical and operational plans. The focus of
this article is the short-term (operational) plan scheduling.
As detailed by Schmidt et al. (2000), the operational level schedules operations to assure
in-time delivery of final products. Here the question is when to perform a manufacturing task
and at which facility so that due dates are met to the fullest extent possible.
In an underground mine context, each production stope has to go through a cycle of
activities until ore is finally produced. Thus, we have to schedule machines to each activity of
each task (production area) in a way that the final production will be the best.
Consequently, a short-term (operational) plan has to handle with a high level of detail,
that is, a greater number of activities compared to tactical or strategic plans. Depending on
mining method and other mine specificities, each mine will plan and monitor its own set of
activities. Next figure presents an example of AngloGold Ashanti Cuiabá mine in Brazil. In
this example, we have the biggest cycle considered, that means, some production areas cycles
can skip one or more of these activities.
The basic, and most common approach, to schedule a short-term plan is to do it manually
in accordance with the planning and production engineers’ experience. Whether done with
Gantt based planning tools or with spreadsheets, this method is normally centered on greedy
algorithms such as anticipating high grade ore or anticipating lowest duration (easier) tasks.
This approach can lead to suboptimum results of the overall production.
© 2019 Taylor & Francis Group, London, ISBN 978-0-367-33604-2
386
Understanding plan’s priorities: Short term scheduling optimization
A.B. Andrade
AngloGold Ashanti, Sabará, MG, Brazil
P.C.B. Rampazzo
Unicamp, Limeira, SP, Brazil
ABSTRACT: Three levels are usually applied to mining planning: Long-term sets the
strategy while medium and short-terms aim to achieve main targets (ore production, grades,
meters developed, metal production, etc.). Manually scheduling a monthly plan, which is a
widely used method, can be a difficult task: A 30-day plan can take days to be elaborated and
appropriately validated. In a dynamic mine environment where day-to-day compliance varies, after the first planned day, some of the tasks will not be executed, some might be delayed
and some anticipated. It will be needed to review the plan and so next day operation can be
correctly prioritized. Most of the time, rescheduling the full plan might not be an option and
suboptimal decisions are made. This paper proposes an algorithm for short-term scheduling
in underground mines with a flow shop scheduling with parallel machines model, optimizing
and automatizing short term planning process.
Keywords: mixed linear programming, short-term scheduling optimization
1 INTRODUCTION
Mining industry, as any capital intensive production system, requires detailed planning to
guarantee that resources will be allocated the best way and that returns on investments will
happen with predictability.
Usually mine planning is carried out in three different levels. Long-term will set the strategic main targets; medium and short-term, the tactical and operational plans. The focus of
this article is the short-term (operational) plan scheduling.
As detailed by Schmidt et al. (2000), the operational level schedules operations to assure
in-time delivery of final products. Here the question is when to perform a manufacturing task
and at which facility so that due dates are met to the fullest extent possible.
In an underground mine context, each production stope has to go through a cycle of
activities until ore is finally produced. Thus, we have to schedule machines to each activity of
each task (production area) in a way that the final production will be the best.
Consequently, a short-term (operational) plan has to handle with a high level of detail,
that is, a greater number of activities compared to tactical or strategic plans. Depending on
mining method and other mine specificities, each mine will plan and monitor its own set of
activities. Next figure presents an example of AngloGold Ashanti Cuiabá mine in Brazil. In
this example, we have the biggest cycle considered, that means, some production areas cycles
can skip one or more of these activities.
The basic, and most common approach, to schedule a short-term plan is to do it manually
in accordance with the planning and production engineers’ experience. Whether done with
Gantt based planning tools or with spreadsheets, this method is normally centered on greedy
algorithms such as anticipating high grade ore or anticipating lowest duration (easier) tasks.
This approach can lead to suboptimum results of the overall production.