Ehsan Ardjmand

Applying genetic algorithm to a new location and routing model of hazardous materials

Nowadays – particularly in systems dealing with hazardous materials (HAZMAT) – in addition to minimising the cost of operations in facility location and routing problems, the risk of these operations is considered an important objective. In this paper, a new mathematical model for the location and routing in facilities and disposal sites is proposed. Also, the risk and cost of transporting goods from facilities to customers is considered. The model minimises weighted sum of the cost and risk by answering these questions: (1) where to open the facilities which produce HAZMAT; (2) where to open disposal sites; (3) to which facilities every customer should be assigned; (4) to which disposal site each facility should be assigned; (5) which route a facility should choose to serve the customers; and (6) which route a facility should choose to reach a disposal site. A novel GA is applied to solve the mathematical model. The results show the robustness of GA in terms of finding high-quality non-dominated solutions and running time.

A robust optimisation model for production planning and pricing under demand uncertainty

The profitability of every manufacturing plant is dependent on its pricing strategy and a production plan to support the customers’ demand. In this paper, a new robust multi-product and multi-period model for planning and pricing is proposed. The demand is considered to be uncertain and price-dependent. Thus, for each price, a range of demands is possible. The unsatisfied demand is considered to be lost and hence, no backlogging is allowed. The objective is to maximise the profit over the planning horizon, which consists of a finite number of periods. To solve the proposed model, a modified unconscious search (US) algorithm is introduced. Several artificial test problems along with a real case implementation of the model in a textile manufacturing plant are used to show the applicability of the model and effectiveness of the US for tackling this problem. The results show that the proposed model can improve the profitability of the plant and the US is able to find high quality solutions in a very short time compared to exact methods.

The discrete Unconscious search and its application to uncapacitated facility location problem

In this paper a discrete variant of Unconscious search (US) for solving uncapacitated facility location problem (UFLP) is proposed. Unconscious search mimics the process of psychoanalytic psychotherapy in which the psychoanalyst tries to access the unconscious of a mental patient to find the root cause his/her problem, which is encapsulated in unconsciousness. Unconscious search is a multi-start metaheuristic which has three main stages, namely construction, construction review and local search. In construction phase a new solution is generated. In construction review the generated solution in construction phase is used to produce more starting points for using in the local search phase. The results of applying US to UFLP shows that this metaheuristic can determine high quality solutions in short processing time comparing to other heuristics.

Water demand forecasting: review of soft computing methods

Demand forecasting plays a vital role in resource management for governments and private companies. Considering the scarcity of water and its inherent constraints, demand management and forecasting in this domain are critically important. Several soft computing techniques have been developed over the last few decades for water demand forecasting. This study focuses on soft computing methods of water consumption forecasting published between 2005 and 2015. These methods include artificial neural networks (ANNs), fuzzy and neuro-fuzzy models, support vector machines, metaheuristics, and system dynamics. Furthermore, it was discussed that while in short-term forecasting, ANNs have been superior in many cases, but it is still very difficult to pick a single method as the overall best. According to the literature, various methods and their hybrids are applied to water demand forecasting. However, it seems soft computing has a lot more to contribute to water demand forecasting. These contribution areas include, but are not limited, to various ANN architectures, unsupervised methods, deep learning, various metaheuristics, and ensemble methods. Moreover, it is found that soft computing methods are mainly used for short-term demand forecasting.

Training the feedforward neural network using unconscious search

One of the most widely used neural networks (NN) is the feedforward neural network (FNN). The most frequent application of FNN is in recognizing nonlinear patterns and, as a nonparametric method, in the estimation of functions especially in forecasting. In this study we will attempt to illustrate how a new metaheuristic algorithm known as Unconscious Search (US) may be utilized to train any feedforward neural network. US operates via a multi-start, memory-based, structured search algorithm that simulates the psychoanalytic psychotherapy process. The Theory of Psychoanalysis, propounded by Sigmund Freud is generally recognized as a descriptive and highly objective account of the mechanisms involved in psychological processes. This paper describes an analogy between the practice of psychoanalysis and the treatment of optimization problems, and it is the task of the present paper to apply US to the problem of training neural network. For this purpose we will first introduce US briefly then an application of US in training FNN is proposed and two benchmark problems are solved and the results of US are compared with the results of other metaheuristic algorithms.

A State-Based Sensitivity Analysis for Distinguishing the Global Importance of Predictor Variables in Artificial Neural Networks

Artificial neural networks (ANNs) are powerful empirical approaches used to model databases with a high degree of accuracy. Despite their recognition as universal approximators, many practitioners are skeptical about adopting their routine usage due to lack of model transparency. To improve the clarity of model prediction and correct the apparent lack of comprehension, researchers have utilized a variety of methodologies to extract the underlying variable relationships within ANNs, such as sensitivity analysis (SA). The theoretical basis of local SA (that predictors are independent and inputs other than variable of interest remain “fixed? at predefined values) is challenged in global SA, where, in addition to altering the attribute of interest, the remaining predictors are varied concurrently across their respective ranges. Here, a regression-based global methodology, state-based sensitivity analysis (SBSA), is proposed for measuring the importance of predictor variables upon a modeled response within ANNs. SBSA was applied to network models of a synthetic database having a defined structure and exhibiting multicollinearity. SBSA achieved the most accurate portrayal of predictor-response relationships (compared to local SA and Connected Weights Analysis), closely approximating the actual variability of the modeled system. From this, it is anticipated that skepticisms concerning the delineation of predictor influences and their uncertainty domains upon a modeled output within ANNs will be curtailed.