ABSTRACT
Combinatorial optimization problems are frequently encountered in a broad range of disciplines from economics to engineering. These problems are notoriously difficult to solve for many applications, especially when the number of free parameters is large. In many situations, high-performance computers are required to solve optimization problems in practical times. Many of the problems solved by neural networks (e.g. pattern recognition and associative memory) can be viewed as special forms of combinatorial optimization. Due to their inherent parallel computation and simple computational requirements, neural networks have the potential to solve difficult optimization problems while taking full advantage of parallel and simple processing hardware. The methodology for applying neural networks to combinatorial optimization problems will be discussed in this section. A brief summary of alternative optimization techniques is also presented.
