Decomposition-based multi-objective distribution estimation optimization method

Abstract

The invention relates to a decomposition-based multi-objective distribution estimation optimization method. The method includes the following steps that: 1) an external population (EP) is initialized to be empty; 2) a group of weight vectors is initialized; 3) the weight vectors are utilized to decompose an original multi-objective optimization problem into a plurality of single-objective optimization sub problems; 4) probability vectors are utilized to model for each decomposed sub problem; 5) each single-objective problem is optimized through randomly sampling the probability vectors, so that new solutions can be generated; and 6) all calculated new solutions are saved to the EP, and whether a termination condition is satisfied is judged, if the termination condition is not satisfied, the method returns to step 3), if the termination condition is satisfied, the method stops, optimal solutions in all the sub problems can be obtained.

Description

technical field [0001] The invention relates to the technical field of combining computer and industrial production, in particular to a multi-objective distribution estimation optimization method based on decomposition. Background technique [0002] Combinatorial Optimization (Combinatorial Optimization) is a branch of Operations Research, which aims to find the optimal solution from discrete or discretizable feasible solutions. Combinatorial optimization problems widely exist in various fields of production and life, including industrial engineering, computer-aided design, computational biology and economic management. These real-world problems are abstracted into different theoretical problems, such as minimum spanning tree problem, knapsack problem, traveling salesman problem, vehicle routing problem, etc. Early research focused on finding optimal algorithms for these problems. However, the development of computational complexity theory has made people realize that some...

AI Technical Summary

Problems solved by technology

But the simple model also has the following two main problems: first, each component in the probability vector is independent of each other, and its use implies the assumption that the variables are independent, so it cannot learn the relationship between variables; second, in When the number of populations is small, the components of the probability vector are easy to converge, which will cause the algorithm to fall into a near-optimal solution

The current research on distribution estimation methods mainly focuses on the design of complex probability models. Relatively complex probability models can improve the ability to solve problems, but their shortcomings are also obvious, mainly reflected in two aspects: firstly, using complex models, Increased computational overhead of the algorithm

Secondly, the use of complex models may weaken the generalization ability of the algorithm, that is, it has a better effect on the problem that the structure and the model match, but the ideal result may not be obtained when the problem structure and the model are different.

Images