Umair F. Siddiqi

Multi-constrained route optimization for Electric Vehicles (EVs) using Particle Swarm Optimization (PSO)

Route optimization (RO) is an important feature of the Electric Vehicles (EVs) which is responsible for finding optimized paths between any source and destination nodes in the road network. In this paper, the RO problem of EVs is solved by using the Multi Constrained Optimal Path (MCOP) approach. The proposed MCOP problem aims to minimize the length of the path and meets constraints on total travelling time, total time delay due to signals, total recharging time, and total recharging cost. The Penalty Function method is used to transform the MCOP problem into unconstrained optimization problem. The unconstrained optimization is performed by using a Particle Swarm Optimization (PSO) based algorithm. The proposed algorithm has innovative methods for finding the velocity of the particles and updating their positions. The performance of the proposed algorithm is compared with two previous heuristics: H_MCOP and Genetic Algorithm (GA). The time of optimization is varied between 1 second (s) and 5s. The proposed algorithm has obtained the minimum value of the objective function in at-least 9.375% more test instances than the GA and H_MCOP

Multi constrained Route Optimization for Electric Vehicles using SimE

Route Optimization (RO) is an important feature of Electric Vehicles (EVs) navigation system. This work performs the RO for EVs using the Multi Constrained Optimal Path (MCOP) problem. The proposed MCOP problem aims to minimize the length of the path and meets constraints on travelling time, time delay due to traffic signals, recharging time and recharging cost. The optimization is performed through a design of Simulated Evolution (SimE) which has innovative goodness, allocation and mutation operations for the route optimization problem. The simulations show that the proposed algorithm has performance almost equal to or better than the Genetic Algorithm (GA) and it requires 0.5N (N is the population size and N ≥ 2 and generally N = 20) times lesser memory than the GA.

A memory efficient stochastic evolution based algorithm for the multi-objective shortest path problem

Multi-objective shortest path (MOSP) problem aims to find the shortest path between a pair of source and a destination nodes in a network. This paper presents a stochastic evolution (StocE) algorithm for solving the MOSP problem. The proposed algorithm is a single-solution-based evolutionary algorithm (EA) with an archive for storing several non-dominant solutions. The solution quality of the proposed algorithm is comparable to the established population-based EAs. In StocE, the solution replaces its bad characteristics as the generations evolve. In the proposed algorithm, different sub-paths are the characteristics of the solution. Using the proposed perturb operation, it eliminates the bad sub-paths from generation to generation. The experiments were conducted on huge real road networks. The proposed algorithm is comparable to well-known single-solution and population-based EAs. The single-solution-based EAs are memory efficient, whereas, the population-based EAs are known for their good solution quality. The performance measures were the solution quality, speed and memory consumption, assessed by the hypervolume (HV) metric, total number of evaluations and memory requirements in megabytes. The HV metric of the proposed algorithm is superior to that of the existing single-solution and population-based EAs. The memory requirements of the proposed algorithm is at least half than the EAs delivering similar solution quality. The proposed algorithms also executes more rapidly than the existing single-solution-based algorithms. The experimental results show that the proposed algorithm is suitable for solving MOSP problems in embedded systems.

Algorithm for parallel inverse halftoning using partitioning of Look-Up Table (LUT)

The look-up table (LUT) method for inverse halftoning is fast and computation-free technique employed to obtain good quality images. In this work we propose a new algorithm to parallelize the LUT method so that more pixels can be concurrently inverse halftoned using minimum additional hardware. The proposed algorithm partitions the single LUT of serial LUT method into N smaller look-up tables (s-LUTs) such that the total number of entries in all s-LUTs remain equal to the number of entries in the single LUT of serial LUT method. The proposed algorithm can be implemented on a single FPGA (field programmable gate arrays) device with external memories to store s-LUTs.

Finding Multi-Objective Shortest Paths Using Memory-Efficient Stochastic Evolution Based Algorithm

Multi-objective shortest path (MOSP) computation is a critical operation in many applications. MOSP problem aims to find optimal paths between source and destination nodes in a network. This paper presents a stochastic evolution (StocE) based algorithm for solving the MOSP problem. The proposed algorithm works on a single solution and is memory efficient than the evolutionary algorithms (EAs) that work on a population of solutions. In the proposed algorithm, different sub-paths in the solution are considered as its characteristics and bad sub paths are replaced by good sub-paths from generation to generation. The proposed algorithm is compared with non-dominated sorting genetic algorithm-II (NSGA-II), micro genetic algorithm (MicroGA), multi-objective simulated annealing (MOSA), and a straight-forward StocE. The comparison results show that the proposed algorithm generally performs better than the other algorithms that works on a single solution (i.e. MOSA and straight-forward StocE) and also infrequently performs better than the algorithms that work on a population of solutions (i.e. NSGA-II and MicroGA). Therefore, our proposed algorithm is suitable to solve MOSP in embedded systems that have a limited amount of memory.

Parallel algorithm for hardware implementation of inverse halftoning

A parallel algorithm and its hardware implementation are proposed for an inverse halftone operation. The algorithm is based on lookup tables from which the inverse halftone value of a pixel is directly determined using a pattern of pixels. A method has been developed that allows accessing more than one value from the lookup table at any time. The lookup table is divided into smaller lookup tables, such that each pattern selected at any time goes to a separate smaller lookup table. The 15-pixel parallel version of the algorithm was tested on sample images and a simple and effective method has been used to overcome quality degradation due to pixel loss in the proposed algorithm. It can provide at least 4 times decrease in lookup table size when compared with a serial lookup table method implemented multiple times for the same number of pixels.

A New Heuristic for the Data Clustering Problem

This paper presents a new heuristic for the data clustering problem. It comprises two parts. The first part is a greedy algorithm, which selects the data points that can act as the centroids of well-separated clusters. The second part is a single-solution-based heuristic, which performs clustering with the objective of optimizing a cluster validity index. Single-solution-based heuristics are memory efficient as compared with population-based heuristics. The proposed heuristic is inspired from evolutionary algorithms (EAs) and consists of five main components: 1) genes; 2) fitness of genes; 3) selection; 4) mutation operation; and 5) diversification. The attributes of the centroids of clusters are considered as genes. The fitness of a gene is a function of two factors: 1) difference between its value and the same attribute of the mean of the data points assigned to its cluster and 2) the frequency with which it has been mutated in previous iterations. The genes that have low fitness values should be updated through the mutation operation. The mutation operation performs small change (positive or negative) in the value of the gene. The mutants are accepted if they are better (with respect to objective function) than their parents. However, diversification in the search process is maintained by allowing, with a small probability, the mutants to replace their parents even they are not better than them. The objective functions used in the proposed heuristic are Calinski Harabasz index and Dunn index. The proposed algorithm has been experimented using real-life numeric data sets of UCI repository. The number of data points and number of attributes in the datasets lie between 150–11 000 and 4–60, respectively. The results indicate that the proposed algorithm performs better than two standard EAs: 1) simulated annealing algorithm and 2) differential evolution algorithm and a genetic algorithm-based clustering method.

A Game Theory Based Post-Processing Method to Enhance VLSI Global Routers

The increase in problem size and complexity of the global routing problems has made it harder for the global routers to produce good results. The global routers employ many different techniques to reach good solutions. However, the results on the recent benchmarks (ISPD 2008) reveal that existing global routers need more enhancements in their designs in-order to improve their solution quality. This work proposes a game theory based algorithm that can enhance the solutions of existing global routers. The proposed algorithm models the rip-up and re-route process as a sequential game in which nets act as players. The set of pure strategies of a net consists of different methods to rip-up and re-route its spanning tree. The nets use mixed strategies in which the probability of any pure strategy is based on the estimation of its likelihood to improve the solution. The performance of the proposed method has been evaluated by using it to enhance the solutions of two excellent global routers namely NCTU-GR 2.0 [1] & BFG-R [2]. The proposed method has been experimented using the ISPD 2008benchmarks and found to be successful in enhancing the total-overflow/wire-length of the existing global routers. On four hard-to-route problems, it has improved the total-overflow of three problems when used with NCTU-GR 2.0 and all four problems when used with BFG-R. It has improved the wire-lengths of all sixteen problems for both NCTU-GR 2.0 and BFG-R. The wire-length of NCTU-GR 2.0 was improved by a value ranging from 35–754 edges and that of BFG-R by a value ranging from 6462–15587 edges. While we demonstrated the potential of GT to enhance results of other heuristics, embedding the discussed technique can help produce better global routers as it will help better traversal of search space, and intelligent decision making.

A hybrid particle swarm optimization for component placement in 3D IC design

This paper deals with a component placement algorithm for 3D IC design. The Particle Swarm Optimization (PSO) is a general purpose stochastic algorithm mimicking the behaviors of particles self-organizing a system. The size of solution space is very large in the 3D component placement problem and it is afraid that the objective function value will be degraded. The Clustering Algorithm (CA) is an efficient initial placement algorithm and this algorithm is used for partitioning the placement problem into clusters with the total pseudo wire-length minimization. PSO is applied to each of the clusters for determining the detailed placement of components with the acceleration as well as the objective function optimization. This hybrid PSO (CA-PSO) is experimentally evaluated against a component placement problem of actual printed wiring board consisting of 217 components and 462 nets and the results show its feasibility.

Simulated evolution (SimE) based embedded system synthesis algorithm for electric circuit units (ECUs)

ECU (Electric Circuit Unit) is a type of embedded system that is used in automobiles to perform different functions. The synthesis process of ECU requires that the hardware should be optimized for cost, power consumption and provides fault tolerance as many applications are related to car safety systems. This paper presents a Simulated Evolution (SimE) based multiobjective optimization algorithm to perform the ECU synthesis. The optimization objectives are: optimizing hardware cost, power consumption and also provides fault tolerance from single faults. The performance of the proposed algorithm is measured and compared with Parallel Re-combinative Simulated Annealing (PRSA) and Genetic Algorithm (GA). The comparison results show that the proposed algorithm has an execution time that is 5.19 and 1.15 times lesser, and cost of the synthesized hardware that is 3.35 and 2.73 times lesser than the PRSA and GA. The power consumption of the PRSA and GA (without fault tolerance) are 0.94 and 0.68 times of the proposed algorithm with fault tolerance.