On intelligent human intelligible routing
Dimitar Lakov, George Kirov · 2002
Presents a human-intelligible approach to the solution of Internet routing problems. This paper suggests a user-friendly dialogue in the definition of customer choices as a compromise between their preferences and temporary Internet abilities. Such a compromise arises in the comparison of customer needs for timely, price-affordable, secure and safe Internet transactions with the multi-variant, address-table-confined, congestion-free and delay-dependent constraints existing in real tasks. These contradictory ill-defined factors are estimated using the fuzzy sets paradigm. System design comprises a combination-and-simulation approach to all possible and timely Internet-available paths as reasonable solutions of a pre-defined fuzzy system representing the best choice. The final decision leads to the immediate starting of an Internet transaction procedure. This implementation resembles a habitual system of Internet dialogue boxes in the case of sole priority definition, but differs from it because of the human-intelligible and flexible features of customer choice. A two-level hierarchical system for a good solution is proposed. On the upper, expert-accessible level, a system structure is defined, while the lower level corresponds to a customer-accessible dialogue system, which realizes a decision compromise. Both levels are built in a Fuzzy-Java shell and are experimented with using models of representative Internet structures.