Arab J Sci Eng DOI 10.1007/s13369-015-1934-0 RESEARCH ARTICLE - ELECTRICAL ENGINEERING Application of a Galaxy-Based Search Algorithm to MIMO System Capacity Optimization Abdelmadjid Recioui 1 Received: 18 June 2014 / Accepted: 19 October 2015 © King Fahd University of Petroleum & Minerals 2015 Abstract The objective of this paper is to use the recently proposed galaxy-based search optimization algorithm to enhance the capacity of a multiple input multiple output (MIMO) system with rectangular arrays at both communica- tion ends (transmitter and receiver). This new optimization tool has been recently introduced and is a metaheuristic tech- nique inspired by the dynamics of galactic arm spirals. It is characterized by its robustness, immunity to local optima trapping, relative fast convergence and ease of implementa- tion. The idea is to extend the results obtained for the one- dimensional array geometry to the two-dimensional case. The purpose is to find out which array geometrical dimen- sions produce the highest capacity value. Compared to the linear array case, promising capacity values are found using the two-dimensional arrays which suggests their deployment in future MIMO communication systems. Keywords MIMO systems · Capacity enhancement · Galaxy-based search algorithm · Two-dimensional antenna arrays 1 Introduction MIMO techniques constitute an area of interest in nowa- days wireless communication. The employment of multiple antennas leads to a significant improvement in capacity with- out any additional allocated power or bandwidth [1]. MIMO B Abdelmadjid Recioui rec79dz2002@yahoo.com; recmad2006@yahoo.fr 1 Laboratory Signals and Systems, Institute of Electrical and Electronic Engineering, University M’hamed Bougara Boumerdes, Avenue de l’indépendance, 35000 Boumerdes, Algeria systems exploit the multipath structure of the propagation channel. However, correlations among channel coefficients are influenced by the antenna properties. As the antennas are collocated in a MIMO array, mutual coupling effects may occur [2]. All these effects should be considered when designing an antenna array for MIMO systems. Many MIMO antenna design optimization studies have been reported in literature. Capacity formulas are given in [3, 4] for mutual coupling and spatial correlation effects with a Rayleigh fading channel being assumed. In addition, the geometries considered constitute namely uniform linear arrays with spacing between the elements being optimized. In [5, 6], the issue of how designers can appropriately select the number of antennas at the asymmetric base station and mobile units has been addressed. In [7], optimizing the MIMO system capacity with unequal costs of implementing antennas at both channel ends has been dealt with. However, in this work, the cost function is expressed using approx- imated asymptotic expression for the ergodic capacity cal- culations. From geometry selection point of view, uniform linear array (ULA) is the most common geometry in mod- ern wireless systems. The uniform circular array (UCA) is as an alternative geometry with some enhanced properties. The results in [8–10] show that the spatial correlation decreases for UCA compared to ULA on average for small and mod- erate angular spread (AS) for similar aperture sizes. On the other hand, ULA has less spatial correlation than UCA for near broadside angle-of-arrivals with moderate AS. Recioui and Bentarzi [11, 12] challenged capacity maximization through optimizing the spacings between the elements of a linear array taking into account mutual coupling and spatial correlation simultaneously. A channel model was proposed to account for both mutual coupling and spatial correlation and has been validated through comparing it with its inde- pendent counterparts. The results reveal an enhancement in 123