Page 314 - Practical Design Ships and Floating Structures

289

decomposed in such as way that the number of linking variables can be minimised. Hence the need for
the algorithm.
TABLE 1
A PARTIAL TABLE SHOWING THE HYPERGRAPH REPRESENTATION OF THE DESIGN PROBLEM

The result of the decomposition exercise is shown in Table 2. The two resultant sub-problems are
semi-independent, and can be solved independently when the three linking variables XI, x3 and x4 are
co-ordinated. Sub-problem one is related to the economic objectives and their associated constraint
functions. Sub-problem two is related to survivability objectives and their associated constraint
functions. The two sub-problems could then be handled by two teams of designers: sub-problem one
by the preliminary ship design team and subproblem two by the safety analysis team. This illustrative
example concentrates on solving subproblem one as some concern is assumed to exist in this area.
The design variables involved in this subproblem are: XI (Length, L), xz (Breadth, B), x3 (Depth, D),
XI (Draught, T), x5 (Speed, V), 16 (Block coefficient, cb) and x7 (Waterplane coefficient C,,,). The
choice of variables involved in the sub-problem obviously depends on the mathematical models used.
TABLE 2
A PARTIAL TABLE SHOWING THE RESULTOF DESIGN DECOMPOSITION EXERCISE

4.2 Design Reuse

As discussed above, a search in a database of good designs for design data that meets the design
specification is the simplest and easiest form of design reuse. For a relatively simple design