Abstract-This paper investigates an assembly sequence planning approach with fireworks algorithm. To solve the problem of assembly sequence planning, the relevant operations of fireworks algorithm are discussed, and the frequency of assembly orientation change, the frequency of assembly tool change, and the frequency of operation change are taken into account in the fitness function of assembly sequence planning. The steps of fireworks algorithm for assembly sequence planning are investigated and finally a case study is given to verify the algorithm.Index Terms-Fireworks algorithm, assembly sequence planning, multi-objective optimization.
I. INTRODUCTIONAssembly planning consists of three parts, including assembly modeling, assembly sequence planning, and assembly path planning. The key of the assembly planning is the assembly sequence planning, which directly affects the assembly cost of the products. As the number of the assembly sequence increases exponentially when the components of the product increase, so the complex products which have a lot of components will cause the combinatorial explosion problem about assembly sequence planning. Therefore, the method of assembly sequence planning will fall into the class of nondeterministic polynomial bounded hard problems [1].The traditional methods [2], [3] are difficult to solve assembly sequence planning for complex products. In recent 20 years, many scholars have applied the intelligent optimization algorithms to the assembly sequence planning, such as genetic algorithm (GA) [4], particle swarm optimization (PSO) [5], simulated annealing algorithm and ant colony algorithm (AC) [6], which provided the powerful methods to obtain the assembly sequence of the complex products. The efficiency of the process to search feasible assembly sequence is proved by the intelligent optimization algorithm.Fireworks algorithm was firstly proposed by Tan and Zhu according to the natural phenomenon of the spark fireworks explosion in 2010 [7]. Fireworks algorithm with the ability of local search and global search has good optimization performance, and the searching mechanism is different with the other algorithms, which makes it searching the solution space more thoroughly. In this paper, to solve the problem of assembly sequence planning, the firework algorithm is applied in assembly sequence planning for the first time, to find the optimal or near-optimal assembly sequence of the product.
II. PRINCIPLE OF FIREWORKS ALGORITHMIn the fireworks algorithm [7], n initial locations are first selected, where n fireworks are set off. After the explosion of the fireworks, the locations of sparks can be obtained, and the quality of the locations can be evaluated. The algorithm should stop when the optimal location is found; otherwise, n current locations should be selected from the fireworks and sparks for the next generation of explosion. Some key technologies of the fireworks algorithm are discussed as follows:where m represents a parameter which controls the total number of sparks gene...