A genetic algorithm to generate and evaluate assembly plans

Abstract

This paper describes a genetic algorithm that deals with the assembly plaizning problem. While iiiost usseriibly plaiiniiig systeiiis use a cut-set iiiethod to generate usseiiibly plans, we propose a new approach to this problem: we use a genetic algorithm that generates and evaluates asseinbly plans. This algorithm starts from a set of valid assenibly plans proposed by an expert of the product. This set is the iiiitinl popirlutiori of poteiitiril solutions. Each asseinbly plan is eiicoded into a chronimonie, to be iiianipulatecl by genetic operators. A reproduction process uses these operntors to prodiice new asseinbly plciiis fi-om "parents" assembly plans. An evalirntioii jiiiiction nricl a selection procedure retniii the best plans that expmid the popirlatioii and sene for new generations. This algorithrii can be used either to quickly generate a set of good assembly plans, or to search all the valid assembly yluiis of a procliict.

Robotic Assembly Sequence Planning and Optimization by Cuckoo Search Algorithm

“…The optimum set of parameters found to be able to provide the optimal/near optimal assembly sequence in reasonable time is population size of 10, fraction of worst eggs that are thrown out 10%, probability used in generating new eggs to replace those that are thrown out 90%. After the CS algorithm is run using the above set of parameters, the optimal assembly sequence obtained is [1,2,3,4,5,6,7,8,9,10,11,12,13,14] with fitness value of 1.5269, requiring 7 tool changes, 4 reorientation changes, stability index of 9, having the base component at the first location of the assembly sequence, and most importantly the sequence was found to have no feasibility (precedence) violations. In order to evaluate the effectiveness of the proposed CS algorithm, it is compared with Improved Harmony Search (IHS) algorithm [12].…”

“…Components [2,5,6] are missing and components [3,4] 3 Robot task level planning for implementation of generated assembly sequence To implement the generated sequence for robotic assembly, it must be converted into a task level plan and executable robot level program. A task level planning strategy for robotic assembly has been developed by a knowledge-based system implemented in Expert System shell, CLIPS [19].…”

“…Hong and Cho [1] developed a robotic assembly sequence optimization approach using Hopfield neural network and expert system to infer assembly constraints from product liaison data. Bonneville et al [2] proposed a GA approach to generate valid and good assembly plans but its performance was slow, and did not guarantee optimum plans. Mishra and Deb [3] developed a knowledge based system for selection of assembly tools, and a GA based approach to determine optimal assembly sequences considering minimum tool changes and assembly direction changes.…”

Formulation and Search of Assembly Sequence Design Spaces for Efficient Use of Assembly Plant Resources for New Products

“…Assemble planning problem presented by Bonneville et al [22] uses genetic algorithm (GA) with the initial population is composed of few assembly plans examined by an expert. The approach of assembly sequence planning problem [23] is a modified version of GA with the search space clustered as families of sequences having similar genetic characteristics.…”

Assembly Sequence Planning for Robotic Automatic Operation

“…Bonneville F at al. use genetic algorithm to generate and evaluate assembly plan [7]. De Lit P at al.…”