| Multi-objective optimization means in the same time to optimize several objectives whichare always conflicting. There are two basic metrics in Portfolio, which are Return and Risk. Ingeneral, High risk high return and low risk low return. However, in reality people want higherreturn and lower risk simultaneously. As a result, Portfolio can be seen as a multi-objectiveoptimization with two objectives that is risk and return.Ant Colony Optimization (ACO) is one of the artificial-intelligence algorithms, which hasalready applied into many combinatorial optimization problems widely and has got excellentresult. However, nowadays, most researches in Portfolio field which use ACO aresingle-objective ones. This paper presents the Pareto front (effective frontier) solution of aadvanced ACO algorithm on the Portfolio. This is a trial and extension of ACO in Portfolio’sapplication.The main tasks are:1. Through review the literature, the current study in Portfolio and ACO are understood.2. The model of Portfolio under ACO is established.3. New inspiration function, self-adaptive pheromone and max-min pheromone range areadded into ACO based on multi-objective optimization.4. Pareto Front which is also called effective frontier is obtained by the advanced ACOalgorithm.5. Simulated annealing algorithm (SAA) based on multi-objective optimization is alsorealized in the paper. In experiment, the two algorithms are given same data to solve thePareto front correspondingly. U-measure is used to analyze the uniformity of the Pareto frontin case of same iteration number and same Pareto solution number.6. The function and effects of parameter on the performance of the algorithm are analyzedsimply at the end of the paper.Portfolio is meaningful in both theoretical research and practical application. ACO isapplied into multi-objective problem, which is a meaningful trail to the expansion on theapplication ofACO. There is some theoretical significance and practical value in this paper. |
PDF Full Text Request