الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 98 |  |  |
| | | from | 98 | | |
Abstract
The ability of professional naval architects, who have a solid foundation in a variety of fundamental and specialized scientific subjects, was historically the only major factor in determining how ships were designed. Whereas the process of designing ships traditionally relies only on the experience of these engineers. Furthermore, the complexity of the many techno-economic criteria, some of which conflict with one another, and the maze of safety standards set forth by national and international regulations that are already in effect and that are foreseen to exist are the main challenges in ship design. Specifically, three stages are normally included in the design of a ship: concept design, contract design, and detailed design. The first phases of design (concept design) should be taking the least time, where the technical aspects of the operational requirements are translated, this is accomplished by striking a balance between owner requirements and shipbuilding rules and regulations, leading to one or more potential designs. Consequently, ship design engineers typically face many challenges throughout the various ship design stages. The most significant of these challenges come in the preliminary phase of designing any ship, as the designer needs a general design that satisfies all the customer’s needs while also being compliant with the standards of international ship classification bodies in the shortest amount of time possible. This thesis demonstrates how machine learning may assist naval architects in making more accurate decisions throughout the conceptual structural design stage, where decisions must be made rapidly and with limited information available. In the ship’s design, machine learning methods stand out from other conventional approaches because they can provide the desired design for any engineering structure in a very short amount of time, and the resulting design is the best an engineer can produce at this stage. Furthermore, the ability to the development of the machine learning algorithm and its use is not only in the design of the ship’s structure but also in the other branches in the shipbuilding industry such as predicting the total resistance of the ship without entering the complex details of the hydrodynamics calculations. Obviously, one of the most crucial elements in the conceptual structural design stage of any ship is the design of the midship section, therefore the midship section scantlings for container ships will be determined using one of the most significant and advanced machine learning techniques, which is called multivariate regression analysis. Specifically, multivariate regression is a technique for determining the linear relationship between various independent variables and various dependent variables.