ELECTRO-OPTICS  LABORATORY

An optical method for synthesizing a binary matrix representing all feasible solutions of a bounded (input length restricted) NP-complete combinatorial problem is presented. After the preparation of this matrix, an optical matrix-vector multiplier can be used in order to multiply the synthesized binary matrix and a grayscale vector representing the weights of the problem. This yields the required solution vector. The synthesis of the binary matrix is based on an efficient algorithm that utilizes a small number of long-vector duplications. These duplications are performed optically by using an optical correlator. Simulations and experimental results are given.