Introduction

This is demo application for "Multi-Objective Optimization Using Membrane Inspired Evolutionary Algorithm".

Algorithm Description

This algorithm works by partitioning decision space into multi-dimensional volumes (membranes). Dimension of the membranes depends on the dimension of decision space. Symbol objects (solution, objective pairs) are initialized in each partition. Reproduction and mutation rules are applied inside each membrane. This will double the symbol object count. Then symbol objects are extracted from each membrane and sorted by non-domination. Half of the symbol objects are then rejected and remaining symbol objects are sent into corresponding membranes. At this point some of the membranes may contain more or less symbol objects than at the start. If a membrane contains more symbol objects; division rule is applied to it. If a membrane contains less symbol objects it is merged with its neighbor.

Use the form below to:

Change the parameters of algorithm.
Select standard MOOP objective to optimize.

Algorithm Parameters

Initialization Method	:
Elite Mate Percentage	:
Mutation Rate	:
Merge/Divide Threshold	:
Hyper Cube Dimension	:
Partition per dimension	:
Symbol objects per membrane	:

Objective Function

Objective to optimize	:
No of Decision Variables (Decision space Dimension)	:
Lower Bound	:
Upper Bound	:
Iterations	:

Execution Status