On algebra of fractions and applications

Nowadays, many decision problems in economical and technical systems can be shown as Linear Fractional Programming (LFP) models. A significant part of the LFP problems consists of Linear Fractional Boole Programming problems. In order to solve the problems as such, we need operations (coordinate operations) which work on fractions with "numerator-numerator, denominator- denominator" principle. In this thesis, coordinate operations which work on fractions with "numerator - numerator, denominator - denominator" priciple have been defined and their geometric and algebraic properties have been examined. Some basic inequalities have been proved. Inversion properties have been examined. A foundation for the heuristic algorithms have been laid by developing sums according to different strategies. Different heuristic algorithms are used in Knapsack problems' solution since Knapsack problem is in NP-complete class. Therefore, in order to solve these problems, new algorithms in accordance with the algebraic substructure in this thesis have been developed by previously written algorithms for 0/1 maximization knapsack problem and 0/1 minimization knapsack problem have been taken into consideration. Finally, some test problems have been created in order to compare the efficiency of these algorithms which had been developed for the solution for these Knapsack problems and then a KNAPSACK Test Problems Library has been created.