Development of Gradient Based Optimization Techniques to Solve the Assignment Problem in High Speed Real time Applications

Abstract

In the recent development in combinatorial optimization, the generalized assignment problem newlineplays an important role. The generalized assignment problem is also termed the many-to-many newlineassignment problem (MMAP) in the broad set of literature. This thesis introduces a gradientdriven newlineinterior point approach to rectify MMAP. Initially, the cardinality constraint detection is newlineexecuted in this technique. At its subordinating stage, a logarithmic barrier function governed newlinegradient descent algorithm is run for deducing a novel outcome for the considered MMAP. An newlineextensive experimental analysis of the proposed work is performed. A comparative study of the newlineproposed work and existing literature is made newline