Professional Documents
Culture Documents
1
ITU Faculty of Naval Architecture and Ocean Engineering
80626 Maslak İstanbul Turkey
Tel 0212 285 64 93
Fax: 0212 285 65 08
e-mail: helvaci@itu.edu.tr
2
ITU Faculty of Naval Architecture and Ocean Engineering
80626 Maslak İstanbul Turkey
Tel 0212 285 64 93
Fax: 0212 285 65 08
e-mail: insel@itu.edu.tr
A KNOWLEDGE BASED SHIP DESIGN METHOD
A design process is a task to define an object to meet the mission requirements and
comply with a set of constraints. Design of an engineering artefact is usually carried out by
an analysis-synthesis-evaluation cycle. In the design of integrated systems, an overall
analysis is usually not possible and the solution is divided into a number of manageable
parts. These parts can then be analysed individually and combined to give an overall
solution. If the individual parts are not fully independent, integration stage is performed in
an iterative method. As ship design involves a wide range of technologies, the overall task
of integration is very demanding. Another difficulty of the ship design task is caused by the
design requirements and constraints that are virtually unique to the ship design. Ship
design applications have been carried out mainly by human expertise and deterministic
analysis techniques. Application of stochastic methods and formalisation of heuristic
methods in ship design have not been widely utilised yet.
1. Background Knowledge
An attempt to improve ship design methodology (Helvacıoğlu, 2001) is attempted by
adopting three disciplines. These are ship design, knowledge based design (KBD) and
expert systems (ES).
The basic ship design aims to determine major ship characteristics affecting cost and
performance (Taggart, 1980). Basic design includes the selection of ship dimensions, hull
form, power, preliminary arrangement of hull, machinery arrangement, and major
structure. Proper selections assure the attainment of the mission requirements such as good
seakeeping performance, manoeuvrability, desired speed, endurance, cargo capacity, and
deadweight. Furthermore, it includes checks and modifications of the achievement of
required cargo handling capability, quarters, subdivision and stability standards, freeboard
and tonnage measurements, usually defined as constraints. A number of major
developments were observed during late 1960’s and 1970’s which had impact on the
general basic design problem. Use of computers in the design was among the most
significant developments. While the computer effected how basic design is performed,
other changes have impacted on what constitutes the basic design problem.
There are three types of logical reasoning; deduction, induction, and abduction. Logical
deduction is the principle mode of reasoning with certainty. This type of reasoning lends
itself to verification and can be readily formalised. The other two modes of reasoning are
induction and abduction. Table 1 describes these three types of reasoning. Induction is the
method to derive the rule from case and result. Hence induction is regarded as the process
in science by which theories are derived from observations of certain phenomena. A third
mode of reasoning is called abduction. Abduction is the derivation of statements about the
world given logical rules and some logical consequences.
Table 1: Three types of reasoning in relation to the syllogism (Coyne et al., 1990)
(1) (2) (3)
Deduction Case + Rule Result
(1) (3) (2)
Induction Case + Result Rule
(2) (3) (1)
Abduction rule + result case
As a consequence of reasoning process, new facts are created, or relations are defined.
Additionally rules can be rewritten by the reasoning process. According to computational
theory, rewrite rules form the basis of a grammar system that defines the syntax of a
language. This similarity between language and design can be used and in analogy.
Components, such as rooms, furniture, stairs form vocabulary, rules form the grammar,
and the design description forms the sentence.
1.3 Expert Systems
The concept of the ES arose in the 1970’s when AI researchers abandoned the quest for
generally intelligent machines and turned instead to the solution of narrowly focused real-
world problems. The area of ESs is a very successful approximate solution to the classic AI
problem of programming intelligence. Feigenbaum has defined an Expert System as “ ... an
intelligent computer program that uses knowledge and interface procedures to solve
problems that are difficult enough to require significant human expertise for their solution”
(Giarratano and Riley, 1994). Expert Systems are collected rules of thumb from human
experience in a computer. Once an expert's knowledge is captured in a computer program,
many copies can be made as required. Expert Systems are based on the idea that rules are
an effective way to tell computers how to do certain kinds of things that people do (Taylor,
1989). An Expert System is based on an extensive body of knowledge about a specific
problem area. In general, this knowledge will be organised as a collection of rules that
allow the system to draw conclusions from the given data.
INFERENCE ENGINE
AGENDA
AUTOMATIC KNOWLEDGE
EXPLANATION FACILITY
ACQUISITION FACILITY
USER INTERFACE
EXPERT/ USER
KNOWLEDGE ENGINEER
CLIPS ES shell is used to code the program. The aim of this process is to achieve goals
starting from cases or facts. This reasoning is called as forward chaining, so this type of
reasoning is adopted to compile the rules and facts of the case. Depth first search strategy
is used to check the related facts and rules to achieve the goals. Facts of case, TEU and
speed, and heuristic rules are utilised in a deductive reasoning to determine ship
dimensions, position of bulkheads, decks, cargo area.
Task Module B is utilised to find minimum required crew number. The facts, engine
power and grosstonnage are acquired from Task Module A. Rules of interpretative
knowledge obtained from regulations are utilised in a deductive process to obtain crew
number. Interpretative knowledge shows the performance of design such as speed of the
ship, minimum required number of crew. Expert System, knowledge based design and ship
design aspects and their utilisation are shown in Figure 3.
Task Module C deals with a component of ship layout, accommodation block as shown in
Figure 3. Number of decks and deck areas are obtained from Task Module A. Required
rooms, room sizes and topological relations between the rooms are handled by ES from the
interviews and rule books. This module outputs decks, rooms on the deck and room sizes.
All these knowledge are coded in CLIPS and the model is used to define the ship in
computer environment.
Ship Object
Third Level of
First Level of Second Level of
Hierarchial
Hierarchial Hierarchial
Decomposition
Decomposition Decomposition
Figure 2 Decomposition
Figure 3 Conceptual Model
References
Alasehir, E., 2000. Ship Database, Final Year Project, Faculty of Naval Arch.& Ocean
Engng., Istanbul Technical University (in Turkish).
Coyne, R. D., Rosenman, M. A., Radford, A. D., Balachandran, M. and Gero, J. S., 1990.
Knowledge-Based Design Systems, Addison-Wesley Publishing Company,
pp.3-456.
Giarratano, J. and Riley, G., 1994. Expert Systems: Principles and Programming, PWS
Publishing Company, pp.1-594.
Helvacıoğlu, Ş., Utilisation of Expert Systems In Container Ship Design, Phd Thesis, 2001
Taggart, R. (Editor), 1980. Ship Design and Construction, SNAME Publication, pp.1-443.
Taylor, W.A., 1989. What Every Engineer Should Know about Artificial Intelligence, The
MIT Press, pp.143-144.