An Advanced Simulation Method (ASM) for Multiple Objective Problems

Abstract

"Proposed is an advanced simulation method (ASM) for multiple objective decision making (MODM) in an interactive optimization mode. It is an improvement over standard interactive simulation based on what if methods (WIM). Both methods are capable of solving MODM problems with numerous alternatives, however, by comparing solutions found with both methods the ASM was shown to be superior to traditional methods. A unique aspect of the ASM is its capability of reverse problem solving (RPS) wherein the decision-maker (DM) starts with a desired solution and works backward to determine the required inputs. In contrast, the WIM gives the results after the DM varies the inputs (a very time consuming process, since in most real problems there is an unlimited set of alternatives). The ASM is user friendly, considers non-linear objectives, allows the units of measurement of the various objectives to be different, and is more efficient than WIMs. A scaling algorithm adjusts for the diversity in impacts caused by MODM problems with heterogeneous objectives while a variable step size algorithm assists in finding the optimal solution. WIMs are the easiest interactive simulation methods to design and use. There are numerous computer supported WIMs. Solutions are found in an interactive process where the user observes the outputs while continuously changing inputs until outputs are satisfactory. If the relationships among objectives and constraints are very complex the user may have difficulty in selecting ideal values for the inputs e.g. given the constraints some selections may not be feasible, changing inputs in attempts to improve one objective may cause other objectives to deteriorate, etc. In WIMs, however, an experienced user has the flexibility to rapidly ascend to the best solution, since the user can immediately select ideal values, if known, for the inputs. "