Shikha Mehta

Memetic Collaborative Filtering Based Recommender System

Web based Decision Support systems like recommendation systems have become effective tools for decision making in the recent past. However the recommender systems employing conventional clustering techniques (KRS) like K-Means for collaborative filtering, suffer from the limitation of getting local optimum results. This paper presents Memetic Recommender System (MRS) based on the collaborative behavior of memes. Memetic Algorithms (MAs) are considered as one of the most successful approaches for combinatorial optimization. MAs are the genetic algorithms which incorporate local search in the evolutionary scheme. We propose a distinctive strategy to perform local search in memetic algorithms. MRS works in 2 phases-In the first phase a model is developed based on Memetic Clustering algorithm and in the second phase trained model is used to predict recommendations for the active user. Rigorous experiments were conducted to prove the decision support and statistical efficacy of MRS visa vis KRS. Results confirmed that the proposed approach yields much better performance as compared to the conventional collaborative filtering recommender system.

Nature-Inspired Algorithms: State-of-Art, Problems and Prospects

Nature-inspired algorithms have gained immense popularity in recent years to tackle hard real world (NP hard and NP complete) problems and solve complex optimization functions whose actual solution doesn’t exist. The paper presents a comprehensive review of 12 nature inspired algorithms. This study provides the researchers with a single platform to analyze the conventional and contemporary nature inspired algorithms in terms of required input parameters, their key evolutionary strategies and application areas. A list of automated toolboxes available for directly evaluating these nature inspired algorithms over numerical optimization problems indicates the need for unified toolbox for all nature inspired algorithms. It also elucidates the users with the minimum and maximum dimensions over which these algorithms have already been evaluated on benchmark test functions. Hence this study would aid the research community to know what all algorithms could be examined for large scale global optimization to overcome the problem of ‘curse of dimensionality’.

Resource provisioning and work flow scheduling in clouds using augmented Shuffled Frog Leaping Algorithm

The on-demand provisioning and resource availability in cloud computing make it ideal for executing scientific workflow applications. An application can start execution with a minimum number of resources and allocate further resources when required. However, workflow scheduling is an NP hard problem and therefore meta-heuristics based solutions have been widely explored for the same. This paper presents an augmented Shuffled Frog Leaping Algorithm (ASFLA) based technique for resource provisioning and workflow scheduling in the Infrastructure as a service (IaaS) cloud environment. The performance of the ASFLA has been compared with the state of art PSO and SFLA algorithms. The efficacy of ASFLA has been assessed over some well-known scientific workflows of varied sizes using a custom Java based simulator. The simulation results show a marked improvement in the performance criteria of achieving minimum execution cost and meeting the schedule deadlines. Meta-heuristic algorithms explored for workflow scheduling in clouds.An improvement proposed to the meta-heuristic algorithms.An augmented variation of Shuffled Frog Leaping Algorithm (ASFLA) formulated.Obtained solutions are execution cost optimal and also meet deadline constraint.ASFLA outperforms Particle Swarm Optimization and SFLA.

Trust aware social context filtering using Shuffled frog leaping algorithm

In the past few years social context filtering (SCF) systems have become trendier to solve the problem of information overload. Conventional SCF approaches utilize preferences of all nearest neighbors to recommend the items. However, in practice preferences of credible peers / true friends with similar interests influence the decision making process. Thus need of trust aware approaches is being increasingly felt. Incorporating users web of trust information though solves the sparsity and cold start problem prevailing in conventional social context filtering techniques but issue of scalability still remains. The work presents Shuffled frog leaping algorithm (SFLA) based SCF approach to develop trust aware system which is capable of handling all the issues addressed above. The approach performs social context modeling using SFLA based clustering. Subsequently, only the trusted neighbors participate in the process of computing most relevant items. Experimental evaluation over Movielens dataset establishes that SFLA based SCF model significantly outperforms conventional K-means approach. Evaluation over Epinions (rating and trust) dataset further substantiates the accuracy of SFLA based trust aware approach over mean absolute error metric.

Enhanced flower pollination algorithm on data clustering

Abstract Nature-inspired algorithms are emerging as most compatible algorithms in obtaining near-optimal solution to complex problems. The ability of meta-heuristic algorithm to obtain global optimization solution largely depends on convergence behavior. In order to enhance convergence capability of latest nature-inspired algorithm i.e. flower pollination algorithm (FPA), a modified version is presented. The performance of modified FPA is tested over clustering application. Algorithm is assessed in contrast to bat algorithm, firefly algorithm, and conventional FPAs on 10 clustering data-sets. Out of 10 data-sets, 8 are derived from pattern recognition and 2 are artificially generated. Clustering results are computed in terms of objective function value and CPU time taken at each run. Run length distribution graphs illustrate the convergence behavior of algorithms. Results indicate that the proposed modified FPA outperforms its counterparts both in terms of attaining best fitness value and reducing the CPU time.

SEVO: Bio-inspired Analytical Tool for Uni-modal and Multimodal Optimization

With the rising success of bio-inspired algorithms to solve combinatorial optimization problems, research is focused towards design of new bio-inspired algorithms or new variants of existing algorithms. To validate the reliability of new algorithms with respect to the existing algorithms, they are tested using benchmark test functions. However existing automated tools with benchmark test problems are limited to genetic and evolutionary algorithms only. Therefore large group of researchers have to repeatedly write the same code for the existing algorithms along with their own proposed version. To address this need, the paper presents a unified swarm and evolutionary optimization (SEVO) tool that automates established algorithms like genetic algorithm (GA), memetic algorithm (MA), particle swarm optimization (PSO) and shuffled frog leaping algorithm (SFLA) over sixteen benchmark test functions with diverse properties. SEVO tool provides a user friendly interface to the users for input parameters, options to simultaneously execute any combinations of algorithms and generate graphs for comparison. To substantiate the effectiveness of SEVO tool, experiments were performed to compare the abilities of GA, MA, PSO and SFLA to attain global minima and speed of convergence. Results establish that convergence rate of SFLA is significantly better than PSO, MA and GA. SFLA also outperformed PSO, MA and GA in attaining global minima. Thus SEVO toolbox may serve as an imperative aid for the bio-inspired research community to perform simulations of the embedded algorithms over varied classes of optimizations test problems with minimum time and effort.

Context aware filtering using social behavior of frogs

Abstract The problem of information overload surfaced with the emergent popularity of dynamic web applications. To tackle this issue, various context awareness approaches have been developed to filter the information. Conventional context aware social filtering techniques predominantly focus on time and location as context of the users. However, another relevant context that of user׳s demographic information is often left out. The paper presents demographic context based filtering using social behavior of frogs. The approach employs shuffled frog leaping algorithm (SFLA) to perform the context modeling and handle the sparsity and scalability issues in social filtering. The work proposes two distinct methodologies to model the demographic context – SFLA based Contextual two dimensional (SC2D) and SFLA based Contextual three dimensional (SC3D) approach. SC2D approach primarily develops a model based on social behavior and subsequently incorporates the personal demographic (occupation, gender, etc.) context to compute the most relevant items. In the SC3D approach, personal demographic context is amalgamated with social behavior to develop the model and thereafter a contextual model is used to generate most relevant items. Experimental studies revealed that SC2D approach is able to reduce the error up to 15% and 8% as compared to MAC2D and GAC2D, respectively, and SC3D approach improves the accuracy upto 31% with respect to MAC3D and upto 26% as compared to GAC3D. Analysis of variance (ANOVA) test results for all approaches corroborate that the differences between the means of SC2D, MAC2D and GAC2D and SC3D, MAC3D and GAC3D are highly significant. These results improve confidence in SFLA as a better optimization algorithm for context aware filtering.

Improved shuffled frog leaping algorithm for continuous optimisation adapted SEVO toolbox

This paper presents improved shuffled frog leaping algorithm ISFLA with controlled random search behaviour. The work proposes adaptation of random solution generation rule with control parameter to manage the direction of search in conventional SFLA. To evaluate the effectiveness of ISFLA, it has been compared with respect to GA, MA, PSO and SFLA for large dimensions-100, 500 and 1,000 over benchmark test problems using SEVO toolbox. Results depict that ISFLA performs considerably better for all benchmark problems. Results also demonstrated the utility and simplicity of SEVO toolbox for simulating new algorithms. ANOVA test substantiated the statistical significance of the obtained results.

Bio-inspired approach to solve chemical equations

Technological advancements in the recent past have changed the state of art teaching and learning styles to computer based education system. Connecting theory with practice is the motto of modern education system. Chemistry is one of the subjects which cannot be learned without practice. Nevertheless, most of the organizations lack assigned chemistry laboratories and apparatus to perform laboratory experiments. To overcome this problem, the paper presents a bio-inspired approach to automate the task of simulating inorganic chemical experiments. The approach involves generation of possible products from the input reactants using Genetic algorithm. Subsequently System of equations method is adopted to balance the reactions. Experimental evaluations established that proposed approach runs with 95% accuracy. Thus presented approach has the competence to equip anyone with the facility to decide and perform any kind of chemical reaction with any set of reactants and get results in a robust manner.

An Evolutionary Approach to Intelligent Planning

With the explosion of information on WWW, planning and decision making has become a tedious task. The huge volume of distributed and heterogeneous information resources and the complexity involved in their coordination and scheduling leads to difficulties in the conception of optimal plans. This paper presents an intelligent planner which uses modified Genetic Algorithm assisted Case Based Reasoning (CBR) to solve the cold start problem faced by CBR systems and generates novel plans. This approach minimizes the need of populating preliminary cases in the CBR systems. The system is capable of generating synchronized optimal plans within the specified constraints. The effectiveness of the approach is demonstrated with the help of case study on e-Travel Planning. Rigorous experiments were performed to generate synchronized plans with one hop and two hops between train and flight modes of transport. Results proved that GA assisted CBR outperforms the traditional CBR significantly in providing the number of optimized plans and solving cold start problem.