Sri Krishna Kumar

Minimisation of supply chain cost with embedded risk using computational intelligence approaches

Global supply chains are vulnerable towards different types of risks and are dynamically expanding with the increase in globalisation. Costs are associated with every risk factor that causes disturbances in the allocation of certain goods at the required place and time, and with the required quality and quantity. In this paper, we consider a multi-echelon global supply chain model, where raw material suppliers, manufacturers, warehouses and markets are located in different countries. The paper first identifies all types of operational risk factors, their expected value and probability of occurrence, and associated additional cost. Based on initial information for the risk factors, optimal decisions regarding the inter-echelon quantity flow in the supply chain are made for a single planning horizon. Then, with the change in the expected value of the risk factors, the intra-echelon shift of flow is determined in order to minimise the total cost and risk factors. Considering the complexity involved with the problem, various computational intelligence techniques such as genetic algorithms, particle swarm optimisation and artificial bee colony are applied in the solution evaluation phase. The results obtained using the developed model illustrate that the ability to react to changes in risk factors offers potential solutions to robust supply chain design.

Supply chain system design integrated with risk pooling

This paper considers the location, production-distribution and inventory system design model for supply chain for determining facility locations and their capacity. Risk pooling effect, for both safety stock and running inventory (RI), have been incorporated in the system to minimize the supply chain cost along with determining facility location and capacity. In order to study the benefit of risk pooling for safety stock and RI two cases have been considered, first when retailers act independently and second when DCs-retailers work jointly. The model is formulated as mixed integer nonlinear problem and divided into two stages. The first stage determines the optimal locations for plants and flow relation between plants-DCs and DCs-retailers. At this stage the problem has been linearized using piece-wise linear function. Second stage enumerates the required capacity of opened plants and DCs. The first stage problem is further divided in two sub-problems using Lagrangean relaxation. First sub-problem determines the flow relation between plants and DCs whereas; second sub-problem determines the DCs- retailers flow. Solution of the sub-problems provides the lower bound for the main problem. Computational results reveal that main problem is within the 8.25% of the lower bound and significant amount of cost reduction can be achieved for safety stock and RI costs when DC-Retailer acts jointly.

Sustainable maritime inventory routing problem with time window constraints

Maritime inventory routing problem is addressed in this paper to satisfy the demand at different ports during the planning horizon. It explores the possibilities of integrating slow steaming policy as mentioned in Kontovas et al. (2011) and Norstad et al. (2011) within ship routing. A mixed integer non-linear programming model is presented considering various scheduling and routing constraints, loading/unloading constraints and vessel capacity constraints. Non-linear equation between fuel consumption and vessel speed has been incorporated to capture the sustainability aspects. Several time window constraints are inculcated in the mathematical model to enhance the service level at each port. Penalty costs are incurred if the ship arrives early before the starting of the time window or if it finishes its operation after the ending of the time window. Costs associated with the violation of time window helps in maintaining a proper port discipline. Now, owing to the inherent complexity of the aforementioned problem, an effective search heuristics named Particle Swarm Optimization for Composite Particle (PSO-CP) is employed. Particle Swarm Optimization Differential Evolution (PSO-DE), Basic PSO and Genetic Algorithm (GA) are used to validate the result obtained from PSO-CP. Computational results provided for different problem instances shows the superiority of PSO-CP over the other algorithms in terms of the solution obtained.

Multi-objective process planning and scheduling using controlled elitist non-dominated sorting genetic algorithm

The integration of process planning and scheduling is considered as a critical component in manufacturing systems. In this paper, a multi-objective approach is used to solve the planning and scheduling problem. Three different objectives considered in this work are minimisation of makespan, machining cost and idle time of machines. To solve this integration problem, we propose an improved controlled elitist non-dominated sorting genetic algorithm (NSGA) to take into account the computational intractability of the problem. An illustrative example and five test cases have been taken to demonstrate the capability of the proposed model. The results confirm that the proposed multi-objective optimisation model gives optimal and robust solutions. A comparative study between proposed algorithm, controlled elitist NSGA and NSGA-II show that proposed algorithm significantly reduces scheduling objectives like makespan, cost and idle time, and is computationally more efficient.

Bulk wheat transportation and storage problem of public distribution system

The novel bulk wheat transportation and storage problem of public distribution system is addressed.A mixed integer non-linear programming model is developed.Hybrid Chemical Reaction Optimization with Tabu Search (CROTS) is proposed.The analysis of the computational results illustrates the effectiveness of the proposed algorithm.The statistical analysis of the algorithms is performed. This research investigates the multi-period multi-modal bulk wheat transportation and storage problem in a two-stage supply chain network of Public Distribution System (PDS). The bulk transportation and storage can significantly curtail the transit and storage losses of food grains, which leads to substantial cost savings. A mixed integer non-linear programming model (MINLP) is developed after studying the Indian wheat supply chain scenario, where the objective is to minimize the transportation, storage and operational cost of the food grain incurred for efficient transfer of wheat from producing states to consuming states. The cost minimization of Indian food grain supply chain is a very complex and challenging problem because of the involvement of the many entities and their constraints such as seasonal procurement, limited scientific storages, varying demand, mode of transportation and vehicle capacity constraints. To address this complex and challenging problem of food grain supply chain, we have proposed the novel variant of Chemical Reaction Optimization (CRO) algorithm which combines the features of CRO and Tabu search (TS) and named it as a hybrid CROTS algorithm (Chemical reaction optimization combined with Tabu Search). The numerous problems with different sizes are solved using the proposed algorithm and obtained results have been compared with CRO. The comparative study reveals that the proposed CROTS algorithm offers a better solution in less computational time than CRO algorithm and the dominance of CROTS algorithm over the CRO algorithm is demonstrated through statistical analysis.

Ontology mapping using description logic and bridging axioms

In the last decade various proposals have been made to promote fruitful and efficient collaboration among small and medium sized enterprises (SMEs) in the form of virtual enterprises (VEs). The success of VEs depends on seamless interoperability of knowledge and data sharing. Ontology implementation is becoming an essential and successful tool for VE operation but commonly ontology mapping is also required to achieve interoperability. The current state of the art in ontology mapping indicates that mapping systems require a great deal of human intervention as mapping brings various types of conflicts and inconsistencies. The ontology mapping method proposed in this paper uses description logic (DL) based bridging axioms between the ontologies. Atomic concept level similarity has been taken as input to establish the complex concepts and roles level mapping. Manufacturing and marketing enterprise ontologies are considered and their mapping has been demonstrated as an example of the proposed mapping process.

A multi-period inventory transportation model for tactical planning of food grain supply chain

The food grain supply chain problem of the Public Distribution System (PDS) of India is addressed in this paper to satisfy the demand of the deficit Indian states. The problem involves the transportation of bulk food grain by capacitated vehicles from surplus states to deficit states through silo storage. A mixed integer non-linear programming (MINLP) model is formulated which seeks to minimize the overall cost including bulk food grain shipment, storage, and operational cost. The model incorporates the novel vehicle preference constraints along with the seasonal procurement, silo storage, vehicle capacity and demand satisfaction restrictions. The management of Indian food grain supply chain network is more intricate and difficult issue due to many uncertain interventions and its chaotic nature. To tackle the aforementioned problem an effective meta-heuristic which based on the strategy of sorting elite ants and pheromone trail updating called Improved Max-Min Ant System (IMMAS) is proposed. The solutions obtained through IMMAS is validated by implementing the Max-Min Ant System (MMAS). A sensitivity analysis has been performed to visualize the effect of model parameters on the solution quality. Finally, the statistical analysis is carried out for confirming the superiority of the proposed algorithm over the other.

An integrated Bayesian-Game theoretic approach for product portfolio planning of a multi-attributed product in a duopolistic market

The focus of this paper is to develop a Bayesian-Game theoretic framework for product portfolio planning problem thereby aiding the manufacturers operating across variety of product industries to offer the right product portfolio set. The problem is modelled for a duopolistic market and the product type considered is characterised by multiple product attributes having varying attribute levels. Initially, feasible product portfolio candidates are generated in terms of combinations of different product attributes and their attribute levels employing the attribute compatibility constraint. Different product portfolio sets thus generated function as different product offering strategies of the two manufacturers. Thereafter, employing the function-based cost-estimating framework and multi-linear regression methodology, manufacturing costs and product premiums, respectively, are estimated for different product portfolios. Utilising the Bayesian risk network, the purchase probabilities are estimated in high, medium and low-risk states for various product portfolios. The purchase probability is made a function of price and functionality. The purchase probabilities thus obtained acts as an input to the final pay-off calculation. Finally, employing these pay-off values, product offering scenarios are populated for the two manufacturers both in equilibrium and non-equilibrium state.

EMPSO-based optimization for inter-temporal multi-product revenue management under salvage consideration

The retail market is governed by customer behavior, demand pattern and inventory replenishment policies. It is also observed that any decision would prove to be full of errors, and objective of enhancing the market share could not be achieved, without inclusion of these factors and policies. While an extensive set of literature exists on single and multi-product dynamic pricing, the issue of liquidation of leftover inventory has so far received scant attention from the researchers of Operations Management community. The current work primarily tries to bridge this research gap by addressing dual objectives of revenue maximization and reduction of salvaging losses. In this paper an inter-temporal dynamic pricing model for multiple products is developed under a market setup with price-sensitive demand. Ideas proposed by [1] and [2] have been taken into account for constructing a revenue structure. The formulated objective function is found to be tractable for deriving prices and procurement quantities of large product portfolios. A multi-objective problem has been devised to handle the optimization of normal and clearance revenue by satisfying several pragmatic constraints. Subsequently, an effective algorithm deriving its traits from Particle Swarm Optimization has been proposed to address this problem. An illustrative example from retail apparel industry has been simulated and solved by the afore-mentioned approach. To validate the model statistical analysis has been carried out and the managerial insights portrayed to reveal the practical complexities involved.

Description logic–based knowledge merging for concrete- and fuzzy-domain ontologies

Enterprises, especially virtual enterprises, are nowadays becoming more knowledge intensive and adopting efficient knowledge management systems to boost their competitiveness. The major challenge for knowledge management for virtual enterprises is to acquire, extract and integrate new knowledge with the existing source. Ontologies have been proved to be one of the best tools for representing knowledge with class, role and other characteristics. It is imperative to accommodate the new knowledge in the current ontologies with logical consistencies as it is tedious and costly to construct new ontologies every time after acquiring new knowledge. This article introduces a mechanism and a process to integrate new knowledge into the current system (ontology). Separate methods have been adopted for fuzzy- and concrete-domain ontologies. The process starts by finding the semantic and structural similarities between the concepts using WordNet and description logic. Description logic–based reasoning is used next to determine the position and relationships between the incoming and existing knowledge. The experimental results provided show the efficacy of the proposed method.