Mangesh Gharote - Academia.edu (original) (raw)

Uploads

Papers by Mangesh Gharote

2020 IEEE International Conference on Services Computing (SCC)

Service organizations need to comply with numerous data regulations to protect and preserve their... more Service organizations need to comply with numerous data regulations to protect and preserve their customers’ privacy. Any misuse of data and privacy breach can affect the organizations’ reputation and brand image. In service delivery scenarios, such as IT support help desk, agents need to access customer data to serve them effectively. This data often includes sensitive and personally identifiable information of the customer. While some amount of data exposure is needed to serve a customer, however, exposure to more data than required could be a threat to an individual’s privacy. Hence, organizations need to design methodologies to ensure customer privacy while achieving minimal cost of operations.In this paper, we propose the Privacy Enabled Task Allocation (PETA) model for assigning customer requests to agents so that the overall cost of operations and data exposure is minimal. Data exposure is minimized by restricting the amount of data exposure per agent and by regulating the assignment of tasks. The PETA problem is modelled as an integer linear program, which is NP-hard. To solve this combinatorial hard problem, we have designed an allocation algorithm based on the linear programming relaxation for finding a quick feasible solution.

CSI Transactions on ICT, 2022

Proceedings of the Genetic and Evolutionary Computation Conference Companion

2020 IEEE International Conference on Services Computing (SCC)

2015 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), 2015

The Human Capital (HC) based performance measurement has become an important aspect for organizat... more The Human Capital (HC) based performance measurement has become an important aspect for organizations to sustain long term growth. The variable selection for Data Envelopment Analysis (DEA) model becomes challenging with the inclusion of HC measures. We propose an automated model discovery method for variable selection. The rankings obtained from the executives act as the guideline for variable selection. We demonstrate this method under different scenarios including the instances where: complete ground truth is known, ground truth is noisy, and oracle ranking obtained by unobserved variables. We discuss in detail the managerial insights for improvement of the inefficient units. This methodology is applied to compare 40 units of an IT service firm.

Reverse logistics refers to the management of theflow of return goodsfrom the point of consumptio... more Reverse logistics refers to the management of theflow of return goodsfrom the point of consumption to the point of origin, for the purpose of recapturing value by either making it re-usable in a market or by proper disposal. Strict environmental regulations and customer demand for green products, coupled with business opportunities in remanufacturing used products have resulted in producers takipg back their products after use. When a returned product reaches the producer, there is an option of remanufacturing the product to meet demand (apartfrom manufacturingfresh products), or stocking it for later use, or disposing it. In this backdrop, we propose periodic review inventoy control policies involving stocking and disposal options. Selectively stocking returned items turns out to be equivalent to manipulating the eflective lead time of remanufacturing of returned items, which is known to influence the total costs. We use simulation to establishgood policies of diflerent types; a single order level (one-parameter) policy and two-parameter policies involving stocking and disposal options. We propose a heuristic for the single parameter disposal policy and suggest a numerical search method for computing the optimal values of two-parameter policies. However, we recommend one-parameter policies as they have shown to be suficiently eflective in performance, and are much easier to implement. Such policies are expected to form a part of the planning systems in organisations in future, due to increased legislation and commercial pressures.

The objective of this study is to explore the relationship between quality and productivity of di... more The objective of this study is to explore the relationship between quality and productivity of different food service providers. Quality is measured in terms of food and service provided by the service provider. We measure the service productivity from customer perspective and compare the different providers on multiple criteria. A case study on different service providers at a large institution was undertaken to illustrate this relationship. We measure and compare service productivity using DEA-super efficiency model and multi-criterion decision making method, TOPSIS; and the results showed a relationship between quality and productivity.

Computers & Industrial Engineering, 2015

ABSTRACT Greedy and optimization approaches used for assigning trainees to software project requi... more ABSTRACT Greedy and optimization approaches used for assigning trainees to software project requirements can give rise to unstable pairs. The presence of unstable pairs can unnecessarily hamper both trainee and project requirement satisfaction. In this paper, we use the theory of stable matching to remove the unstable pairs. We propose an efficient method to predict preferences for both trainees and project requirements using the utility theory. Our computational experiments suggest that the stable matching approach performs better with average trainee and project requirement satisfaction in terms of the preference ranks of allocated choices. Compared to cost based optimization model, the proposed approach obtains significantly better trainee project match at the expense of small additional allocation cost.

Ordering queries within a workload and ordering joins in a query are important problems in databa... more Ordering queries within a workload and ordering joins in a query are important problems in databases [1]. We give algorithms for the query sequencing problem that scale (small space) and are efficient (low runtime) as compared to earlier work [4]. The errors are small in practice and we are able to further reduce them using geometric repair. We provide a computational comparison of TSP solvers and show extensive testing on benchmark datasets [25] observing its connection to these ordering problems.

The Giant Metrewave Radio Telescope, commonly known as GMRT, is the world's largest low frequ... more The Giant Metrewave Radio Telescope, commonly known as GMRT, is the world's largest low frequency radio telescope. Being a rare resource, efficient planning of GMRT is needed. Planning consists of scheduling a pre-selected large collection of scientific proposals up to half a year in advance. Scheduling becomes complex due to the variety of astronomical, equipment and logistical constraints. The goal

2020 IEEE International Conference on Services Computing (SCC)

Service organizations need to comply with numerous data regulations to protect and preserve their... more Service organizations need to comply with numerous data regulations to protect and preserve their customers’ privacy. Any misuse of data and privacy breach can affect the organizations’ reputation and brand image. In service delivery scenarios, such as IT support help desk, agents need to access customer data to serve them effectively. This data often includes sensitive and personally identifiable information of the customer. While some amount of data exposure is needed to serve a customer, however, exposure to more data than required could be a threat to an individual’s privacy. Hence, organizations need to design methodologies to ensure customer privacy while achieving minimal cost of operations.In this paper, we propose the Privacy Enabled Task Allocation (PETA) model for assigning customer requests to agents so that the overall cost of operations and data exposure is minimal. Data exposure is minimized by restricting the amount of data exposure per agent and by regulating the assignment of tasks. The PETA problem is modelled as an integer linear program, which is NP-hard. To solve this combinatorial hard problem, we have designed an allocation algorithm based on the linear programming relaxation for finding a quick feasible solution.

CSI Transactions on ICT, 2022

Proceedings of the Genetic and Evolutionary Computation Conference Companion

2020 IEEE International Conference on Services Computing (SCC)

2015 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), 2015

The Human Capital (HC) based performance measurement has become an important aspect for organizat... more The Human Capital (HC) based performance measurement has become an important aspect for organizations to sustain long term growth. The variable selection for Data Envelopment Analysis (DEA) model becomes challenging with the inclusion of HC measures. We propose an automated model discovery method for variable selection. The rankings obtained from the executives act as the guideline for variable selection. We demonstrate this method under different scenarios including the instances where: complete ground truth is known, ground truth is noisy, and oracle ranking obtained by unobserved variables. We discuss in detail the managerial insights for improvement of the inefficient units. This methodology is applied to compare 40 units of an IT service firm.

Reverse logistics refers to the management of theflow of return goodsfrom the point of consumptio... more Reverse logistics refers to the management of theflow of return goodsfrom the point of consumption to the point of origin, for the purpose of recapturing value by either making it re-usable in a market or by proper disposal. Strict environmental regulations and customer demand for green products, coupled with business opportunities in remanufacturing used products have resulted in producers takipg back their products after use. When a returned product reaches the producer, there is an option of remanufacturing the product to meet demand (apartfrom manufacturingfresh products), or stocking it for later use, or disposing it. In this backdrop, we propose periodic review inventoy control policies involving stocking and disposal options. Selectively stocking returned items turns out to be equivalent to manipulating the eflective lead time of remanufacturing of returned items, which is known to influence the total costs. We use simulation to establishgood policies of diflerent types; a single order level (one-parameter) policy and two-parameter policies involving stocking and disposal options. We propose a heuristic for the single parameter disposal policy and suggest a numerical search method for computing the optimal values of two-parameter policies. However, we recommend one-parameter policies as they have shown to be suficiently eflective in performance, and are much easier to implement. Such policies are expected to form a part of the planning systems in organisations in future, due to increased legislation and commercial pressures.

The objective of this study is to explore the relationship between quality and productivity of di... more The objective of this study is to explore the relationship between quality and productivity of different food service providers. Quality is measured in terms of food and service provided by the service provider. We measure the service productivity from customer perspective and compare the different providers on multiple criteria. A case study on different service providers at a large institution was undertaken to illustrate this relationship. We measure and compare service productivity using DEA-super efficiency model and multi-criterion decision making method, TOPSIS; and the results showed a relationship between quality and productivity.

Computers & Industrial Engineering, 2015

ABSTRACT Greedy and optimization approaches used for assigning trainees to software project requi... more ABSTRACT Greedy and optimization approaches used for assigning trainees to software project requirements can give rise to unstable pairs. The presence of unstable pairs can unnecessarily hamper both trainee and project requirement satisfaction. In this paper, we use the theory of stable matching to remove the unstable pairs. We propose an efficient method to predict preferences for both trainees and project requirements using the utility theory. Our computational experiments suggest that the stable matching approach performs better with average trainee and project requirement satisfaction in terms of the preference ranks of allocated choices. Compared to cost based optimization model, the proposed approach obtains significantly better trainee project match at the expense of small additional allocation cost.

Ordering queries within a workload and ordering joins in a query are important problems in databa... more Ordering queries within a workload and ordering joins in a query are important problems in databases [1]. We give algorithms for the query sequencing problem that scale (small space) and are efficient (low runtime) as compared to earlier work [4]. The errors are small in practice and we are able to further reduce them using geometric repair. We provide a computational comparison of TSP solvers and show extensive testing on benchmark datasets [25] observing its connection to these ordering problems.

The Giant Metrewave Radio Telescope, commonly known as GMRT, is the world's largest low frequ... more The Giant Metrewave Radio Telescope, commonly known as GMRT, is the world's largest low frequency radio telescope. Being a rare resource, efficient planning of GMRT is needed. Planning consists of scheduling a pre-selected large collection of scientific proposals up to half a year in advance. Scheduling becomes complex due to the variety of astronomical, equipment and logistical constraints. The goal