+91 999-999-9999

Lorem Ipsum is simply dummy text of the printing and typesetting industry.

Request A Free Quote
Name *

Please type your full name.
E-mail *

Invalid email address.
Phone *

Invalid Input
Subject *

Invalid Input

Invalid Input

The Optimization Engine for Demand Forecast system helps users to establish the demand for all the SKUs for the subsequent months.

The demand is categorized demographically as well. This will enable decision makers to plan the sourcing of the Raw Material, and production from the production Lines.

In the absence of a reliable demand forecast, organizations need to maintain high inventory of both raw material and finished goods to account for varying demands. A High inventory of Spares & components becomes imminent as the lead-time for the raw material is unpredictable and procurement of the Spare Parts at short notice may result in delays. Finished goods inventory is important to meet the customer service level and to ensure that there is no shortage of supply. A good demand forecasting system will obviate the need for huge inventory levels, as it will help in planning of production and raw material procurement dynamically.

With an accurate demand forecast, various dynamically taken decisions can become far more accurate. Decisions like when to procure, how much to procure and from who to procure can reach close to optimal ensuring that there is minimal level of inventory at any point of time. There is no need to maintain high levels of raw material and finished goods inventory, and at the same time, the customer service levels can be improved upon.

Based on the previous Demand Data, the software will make an Automatic Forecast for the orders to a high degree of accuracy. The Forecast helps in accounting for the expected orders and reduces the risk due to fluctuating demand patterns without needing to maintain large inventory levels. This could save the organizations from situations like idling capacities followed by large number of orders or excess produce followed by small number of orders. The demand forecast module is based on advanced modeling techniques, which can represent the highly dynamic and non-linear nature of the demand pattern.

Most current forecast models are based on linear modeling techniques, whereas the market behavior is far from linear in most cases. Linear models are not capable of predicting long-term results because of their short-term memory. The predicted results could be determined using a simple linear relation as following.

Demand April = Demand March + Constant With such models, the actual results vary largely from the predicted. The software consists of several heuristics to incorporate the non-linear nature of the market. It is based on Recurrent Neural Network based modeling technique, which is capable of representing a highly non-linear market behavior. Neural Networks have a very long-term memory and can be trained using a large time-series market data. The forecast done using a Neural Network Model is based on long-term market trends and are fairly close to the actual Demand.

Our Forecast is based on advanced non-parametric modeling techniques, which are capable of finding the mapping between the demand and the governing inputs. This relation is automatically established from the data and no prior knowledge of the actual system is required for ascertaining this relation. The technique deployed is different from the conventional parametric and parametric modeling techniques, where the user needs to know the underlying structure of the system a-priori. This is a complete black-box modeling technique. The following figure shows the market forecast data in the highlighted region for a product.

Sourcing Optimization

Stock Allocation deals with optimal sourcing of raw material and other spare parts from the production units and the vendors. Based on the available stock level of the material with the vendor or the production unit, the production capacity, the lead time for supply, the shipping cost, the excise duty for the transport (if the shipment is across the state), etc. the software computes an optimal plan for sourcing of the raw material. The plan generated honors the constraints like availability of the stocks, Manufactures portfolio, lead time etc. and at the same time generates an optimal plan ensuring minimal cost of the shipment.

The delivered Solution will contain the following:

  • It will automate the process of selection of the best vendor in an optimal manner once it is known that deliveries will be delayed or cancelled from another vendor on whom orders had been placed. How much to procure from the Vendor selected and in what quantity I also suggested by the software.
  • The optimization of vendor selection will take into account not only multiple constraints (including Last Dates for Licenses) but also multiple objectives (e.g. cost minimization, delivery service level maximization, etc.)
  • Activate automated alarms for the user on deliveries / dispatches being delayed by any vendor. This will enable user to continuously monitor situations on an exceptional basis and take action whenever required.

In order to ensure the full effectiveness of the above solution, the following information will be required to be collected from the vendors on a regular basis:

  • Dispatch data: if deliveries are going to be delayed, information on this as well as the next expected date of delivery.
  • Schedule on excess capacity available with vendors.

The above could be obtained directly from vendors by web enabling the above solution and providing them the facility to enter data directly. Since the data will not be of a sensitive nature as far as the vendor is concerned he should not find it difficult to enter it directly. Alternately, your field Purchase staff could regularly enter the data.

The software employs a range of well known as well as proprietary algorithms to automatically generate the shipping plan and ensure optimality. For more complex problems where multiple constraints have to be optimized, the software employs techniques like Genetic Algorithm, which has the capability of generating near- optimal solutions in a short period of time while accounting for a large number of constraints.

Genetic Algorithm involves generating a population of feasible solutions, measuring their fitness and selecting solutions for crossover or mutation to produce new solutions for the next population. This process gradually transforms the population and the solutions will converge to the near optimal.

Integration With SAP

  • The data will be synchronized between the SAP application thru BAPI's and Optimization Engine at the time of Production Plan generation (either automatically or when a user runs a production plan).
  • There will not be any duplication of masters in the two Applications. Optimization Engine will use all the masters existing in SAP. The new set of masters, which are not present in SAP, will reside in Optimization Engine.
  • The unsuccessful transactions will be stored in a log file and can be retrieved by the Administrator.
  • The Administrator can roll back a set of transactions made and redo the operation.

Hardware/ Software Required

  • Hardware: dedicated Pentium IV with 256mb RAM and above will be required the optimization Engine.
  • Software: No Additional Software will be required for the standard version of the software. However for the Web-based version of the software, IIS Server and Windows NT/ 2000 will be required.