Manufacturing algebra. Part I: modeling principles
and case study
Enrico Canuto
1
, Manuela De Maddis
2
, Suela Ruffa
2
1
Dipartimento di Automatica e Informatica,
2
Dipartimento di Ingegneria Gestionale e della Produzione
Politecnico di Torino
Torino, 1 1029 Italy
{enrico.canuto, manuela.demaddis, suela.ruffa}@polito.it
Abstract - Manufacturing Algebra provides a set of
mathematical entities together with composition rules, that are
conceived for modeling and controlling a manufacturing system.
In this first paper, only the modeling capabilities are outlined
together with a simple case study. Though the algebra is formally
introduced, the scope of the paper is to familiarize the reader
with the proposed methodology, and to highlight some
peculiarities. To this end formulation is reduced to a minimum
and no theorem is included. Among the algebra peculiarities,
both manufacturing process and the factory layout are neatly
defined in their basic elements, and the link between them is
given. A manufacturing model (parts, operations) need to include
time and space coordinates in order it could be employed by
factory elements like Production Units and Control Units. This
asks for the definition of event and event sequence and of the
relevant discrete-event elements and operators. A further
peculiarity to be clarified in the second part, is the capability of
aggregating algebra elements into higher level components, thus
favoring hierarchical description and control of manufacturing
systems.
Index Terms -Manufacturing Algebra, discrete-event, state
equations, manufacturing systems, modelling.
I. I. INTRODUCTION
The paper deals with the problem of modelling and
controlling manufacturing systems. The proposed method is
based on the mathematical framework offered by the
Manufacturing Algebra (MA) and on discrete-event state
equations as outlined in [1] and [2] . Manufacturing Algebra
has been developed at the end of nineties, and since then for
different reasons, .developments and applications have slowed
down. This conference may be an occasion to review the
algebra fundamentals, to test them through a simple case study
in view of applications e.g. to the automotive industry. The
algebra has been developed to model dynamics of production
processes taking place in factories using mathematical objects
and operations, at different degrees of accuracy and detail. It
is conceived around a few concepts:
1) Direct relation with manufacturing. Algebra elements and
their semantics are related with concepts and objects
found in industrial manufacturing systems (Section III).
2) Separation principle. While manufacturing model
describes the materials flows and the operations that
occur during the manufacturing process of each product,
factory model describes the physical layout, accounting
for machines, transport means and storage places. Both
models are kept as separate entities, but their link is
defined and formulated (Section V.A)
3) Aggregation principle. The principle states that lower-
level entities can be combined to provide a reduced
number of higher-level entities having the same
properties of the lower-level.
4) Multilayer modeling. The manufacturing process is
described at different levels of detail in an incremental
way, based on the aggregation principle.
5) Hierarchical control. Control units are part of the model
of the manufacturing system and are compatible with the
multilayer architecture of the model. The result is a
hierarchical control scheme.
Previous work on manufacturing algebra has been already
published, see [3] , [4] and [5] . Here, the main concepts of the
algebra are recalled and referred to a simple case study, by
reducing the formal burden and avoiding formal proofs. The
following problems are treated: how to describe a
manufacturing process (Section IV) and how to link this
description with the factory layout (Section V). How to
simplify models using the aggregation principle, how to
obtain a multilayer model, how to design control strategies are
treated in a companion paper [19] .
Several methods for modelling and controlling
manufacturing systems have been proposed. A comprehensive
discussion is in [6] . Several textbooks are available: relations
to and discrepancies from [7] will be outlined. The need for a
flexible, self-programmable, closed-loop and distributed
technique to design and implement complex manufacturing
operations seems still an open problem [8] . Proposed
solutions are usually in the form of software architectures and
packages (for simulation, monitoring and control) [9] ,
empirical approaches [10] or network structures [11] , and
usually do not stem from a formal and comprehensive
mathematical background, unless they are based on well
known formal methods such as Petri Nets [12] , or
approximate decomposition techniques and Markov models.
Problems of distributed scheduling [13] , parts routing [14] ,
service rate selection have been extensively studied, though
explicit solutions can be hardly achieved [15] [16] [17] . Since
1536 978-1-4673-1277-6/12/$31.00 ©2012 IEEE
Proceedings of 2012 IEEE
International Conference on Mechatronics and Automation
August 5 - 8, Chengdu, China