【Objective】

2

Undersea optical cable plays a leading role in international communication. In the planning of undersea optical cable communication system

several groups of possible schemes are often determined according to experience. The scheme with relatively better transmission performance and investment cost is then selected. However

with the development of communication equipment and communication technology

the number of optional modulation formats and optical fibers keeps are increasing. The number of alternative schemes is too large to quickly determine the optimal scheme. Therefore

it is necessary to establish a cost optimization model of undersea optical cable communication system to solve the cost optimal scheme.

【Methods】

2

Aiming at the problem of system planning and configuration in the undersea optical cable communication system

a cost optimization model considering the parameters of transmission performance

transmission capacity and supply voltage is established by taking the span length

optical fiber type

number of optical fiber pairs

repeater output power

modulation format and number of practical channels as the variables. The cascaded-lexicographic method is used to solve the cost optimization model.

【Results】

2

The results show that the difference between the planning results of the proposed optimization algorithm and full traversal algorithm are no more than 0.34%. However

the number of operations of the optimization algorithm proposed in this paper is far less than that of the full traversal algorithm

with only 1/100 of that of the full traversal algorithm. Secondly

the number of operations of the cascade genelexicography method is less affected by the range of variables than that of the total traversal method. Finally

by comparing the optimal cost of the undersea cable communication system under different span length

it is found that the planning results have no regularity.

【Conclusion】

2

The variables always influence each other. The optimal cost design of the system cannot be obtained according to experience

so it needs to establish a model. The algorithm used in this paper is not only accurate

but also uses fewer operation times.