The aim of the scientific message is informing of a scientific public on the new scientific direction, based by the author. This direction leans on both to area of reinsurance, and to area of information technologies. It can be correctly named "the methods of the operative rational reinsurance of the especially serious risk based on the evolutionary algorithms".

Let´s consider the following economic problem. Let, for example, an insurance group with reinsurer includes together five members. Let, in particular, a is insurer, b, c, d, e are reinsurers. Also, let SSO is a sum of the insurance responsibility of the everyone subsequent reinsurer, being a part of conditions of the insurance contract, depends on a "rating" of the previous reinsurers - in eyes of everyone subsequent reinsurer:

for the first variant of reinsurance

for the second variant of reinsurance

for the third variant of reinsurance

and so on,

where   are the sums of the planned insurance responsibility by reinsurers c, b, d, e accordingly,   are some functions reflecting "rating" of the reinsurers on the relation to each other and known for them.

The ratings of the reinsurers, expressed by the mentioned functions, can carry objective or subjective character.

It is important, that at the significant number of the members of insurance group the efficiency of acceptance of the decisions is rather urgent.

The considered difficult task can be shown to the modified traveling salesman problem having distinguished economic sense in view of a subject domain of reinsurance.

It is necessary to note, that total of variants of ways of detour in the classical traveling salesman problem represents . And n is the number of points of the detour (number of towns).

From positions of the graph theory a classical traveling salesman problem is known as a task of search of the minimal length of the Hamilton cycle.

In particular,10!=3628800;12! = 479001600 . In case of a total search of variants at n=15 and more points the given task is unsoluble even for one year (!) at computer calculations.

Also at the significant amount of towns frequently applies a method of branches and borders promoting reduction of amount of variants of the search. However, the given method does not guarantee, that during the decision of a task will not be made total search of variants.

The task, considered in the message, is characterized by distinguished criterion function

,

where x_k^safe is the required acceptable sum of the insurance responsibility of the insurer,

S is the sum, on which is insured risk by the insurer a at the subsequent reinsurance,

q is the number of the edge according with the order of detour of the graph, in which the all n members of the insurance group are represented by tops of the graph,

SSO_q is the sum of the insurance responsibility of the q-th member of the insurance group at approach of the insurance occasion, , .

The using of this criterion is caused by characteristic situations, in which the insurance company not aspires to a significant prize at reinsurance of a serious risk so much, how many does not wish to go bankrupt owing to the insurance occasion.

It is represented to perspective exertion of evolutionary (genetic, ant algorithms and other) algorithms of functional optimization for the decision of similar tasks connected to operative search of rational variants of reinsurance of the especially serious risk. Result is the creation of the new scientific direction on a joint of the realm of reinsurance and the realm of information technologies, namely: methods of operative rational reinsurance of especially serious risk on the basis of the evolutionary algorithms.

References

1. Kamynkina M.G., Solntseva E.E. Reinsurance. Practical guide for insurance companies. - Moscow, 1994.
2. Holland J. Genetic algorithms // In a Science World. 1992. N09-10.
3. Insurance. 2008. N11.
4. Bonavear F., Dorigo M. Swarm Intelligence: from Natural to Artificial Systems. - Oxford university Press, 1999.