PLANOPT optimization algorithm is a research product. It is a robust hybrid proprietary optimization algorithm. It came into existence in 1995 when two University Professors working together in the areas of VLSI Layout design and  Multi-disciplinary Facility Layout optimization developed a marvelous technique that could solve the Unequal Area QAP better than all published algorithms. They realized that the results were amazing. No researcher in the field could imagine (at that time) that it is at all possible to solve such a hard optimization problem better than all GA's, SA's and the conventional analytical or heuristic techniques. They decided not to publish the algorithm but to present the algorithm anonymously in the form of a software package as a challenge to other researchers. This is how PLANOPT software package came into existence. Since then, PLANOPT optimization algorithm has been a challenge to the researchers in the field. It becomes obvious from the review of published literature that researchers mostly shied away from comparing their techniques with PLANOPT using randomly generated benchmarks. In fact, published optimization algorithms don't even come close to PLANOPT algorithm.  

The optimization algorithm of  VIP-PLANOPT 2006 program is now improved and advanced. It is based on a hybrid smart growth technique. It generates high quality solutions for large scale problems with minimal computational cost. This is due to the algorithm's embedded optimization philosophy of natural constructive growth while identifying, for each module, the feasible design space with the highest probability of local optima . The design space is then mapped onto a straight line. A pseudo-exhaustive search is then carried out for the optimum solution at each stage of a multi-stage optimization process. VIP-PLANOPT 2006 has specifically introduced the following improvements to the algorithm to reinforce its superiority over all known algorithms:

1) Double precision arithmetic has been introduced to control the instability of the algorithm that was observed in certain cases.

2) New techniques have been introduced to reduce the dependence on the user-specified starting seed that is used to generate random numbers for starting the optimization process.

3) Penalty functions have been introduced for obtaining optimal layouts within  user-specified boundary of simple rectangular shape  or complex composite shapes.