VARIABLE POPULATION MOPSO APPLIED TO MEDICAL VISITS

Abstract

Multi-objective optimization techniques are the ideal support tools for the decision-making process. They provide a set of optimal solutions for each of the significant aspects of the problem, thus summarizing the alternatives to be considered. Having a limited number of alternatives makes it easier for decision makers to perform their tasks, since they can focus their efforts towards the analysis of the available options.In this paper, the main characteristics of multi-objective optimization are summarized, and a real experience is described regarding the optimization of mobile units assignment at a health care company in Argentina using a new method based on swarm intelligence called varMOPSO.Keywords: evolutionary computation, swarm intelligence, particle swarm optimization, multi-objective function optimizationJEL Classification: C61, C65(ProQuest: ... denotes formula omitted.)1. INTRODUCTIONWhen facing an optimization problem, the criteria to optimize (fitness functions) must be defined. If there is only one criterion to optimize, the process is called mono-objective optimization. An example of mono-objective optimization would be buying a car based exclusively on its price. In this example, the optimal solution is the cheapest vehicle.Mono-objective problems are widely studied in the literature, and most searchand optimization-related scientific research works are developed based on problems with a single fitness function.In general, real-world problems require the simultaneous optimization of several criteria. Following the previous example, the buyer might want to minimize car price and maximize available equipment. It can be clearly seen that these two objectives are opposite - the lower the price, the less equipment and vice- versa. In this case, the challenge of the optimization problem is finding the set of solutions that optimize both criteria at once - those cars with the best levels of equipment given the various prices, or, in other words, those vehicles that, for each level of equipment, are the cheapest.Optimization problems are present in every area. This is particularly so in economics, which is based on the optimal utilization of limited resources for multiple and unlimited uses; optimization is one of the central tools for this science, as well as one of the simplest tasks of nature. Finance does not stray from this concept, which is the most important one in this area - the optimization of money through the use of this resource in various investment instruments in order to maximize utilities.Understanding the importance of optimization techniques, this article describes a new optimization method that has been applied to the optimization of a realworld problem. The purpose of this paper is showing the various aspects to be considered when implementing this type of techniques by using a specific case as an example.2. PROBLEM TO SOLVEThe real-world case to solve is the automated assignment of mobile units to medical services in an emergency health care provider in Argentina.The main purpose of this type of companies is to provide pre-hospital health care services of various complexity levels: medical emergencies, urgencies, home medical visits, and scheduled transfer of patients to health care centers.These companies provide medical services for the timely and efficient health care of patients who are ill or have sustained some type of wound and/or injury. They are the second link in the emergency chain, the first one being the family or person accompanying the patient when the need for medical care arises, and the third one being the hospital or health care center and its facilities.Each incident that is reported to this type of companies is classified based on its severity and some of the following categories:* Red: Imminent risk of death.* Yellow: Serious emergency, no risk of death* Green: Home medical visit, low severity event. …