Base Model With Sales Graph and Financial Reserves Analysis

Part 1: Basemodel

In the first bit of the experiment, there was a need to add a graph of sales to the base model for Data Analysis Dissertation Help. This demanded running the simulation experiments while reporting on the sales as well as the financial reserves for the base cover that run for 10 years. A number of considerations were put in place such as whether randomness is to be considered or not. At the same time, the simulation process put into consideration the sensitivity analysis in the AnyLogic model and considering the fact that Profit Margin need to be uniformly distributed in a range of 0.5 and 1%. In the first place, the experiment looked at two areas, which include sensitivity analysis and randomness as far as the simulation model development. The model tool development that was considered includes the AnyLogic simulation software under the support of the AnyLogic Company, which was formerly known as the XJ Technologies.

The consideration of Anylogic simulation is thought some of the simulation methodologies as significant and deploys such platforms like Agent Based Modelling, Discrete Event and System Dynamics (Bazan and German 2012). The Based model looks at the advantages that would be gained from the object oriented model design paradigm and the Java programming language. The use of AnyLogic avails the build-in functionality, which relies on the Java OptQuest. In constructing the model, general parameters shall be put into consideration and the modelling process relies on minutes as the time units. Secondly, the simulation process depends on the real calendar for the purposes of developing the representative models. The simulation model runs up to 6 months.

First, in observing randomness, the experiment focuses on the stochastic and uncertainty evaluated against the capacity of the random number generator. It ought to be noted that the probability distribution functions in the AnyLogic Simulation work by default. This is not limited to the random transitions, random events, and the random layouts, process modelling library blocks, AnyLogic simulation engine and the relevant networks. It is worth noting that Randomness across the AnyLogic simulation process substantially relies on the random number generator which appears by default. In this context, the context will perform a number of processes (Fu-gui et al. 2012). First, the experiment substitutes the default RNG with the required RNG. It is also more recommendable to have a range of the RNGs which shall produce the probability distribution function.

Preparation of the custom RNG calls for the creation of the subclass which can be denoted as the java.util.random. This paves way for the subsequent selection of the randomness section while tapping into the appropriate expression that would return the RNG. In Figure 1, the initialization can be achieved through the default RNG which occurs in the course of the initialized experiment and the simulation run. At any point in time, the simulation should call setDeafultRandomGenerator (Random r), which is also the calling function noted in the base model (Faria and Reis 2015).

Precautions would float the fact that everyone should be aware of running the simulation while setting up the General page, which defines the experiment properties. In case of a more certain call, then it is good to initialize an instance where the variable myRNG forms the subclass. Then the experiment is ready for the calling the probability distribution function as the last parameter. For instance, the calling should carry the property of uniform (myRNG) or sometimes make use of the triangular (5, 10, 25, myRNG). The second bit of the experiment narrows down to the sensitivity analysis while setting the profit margin in range of 5-10% (0.05-0.1). First, the experiment focused on what kind of a variable that needs to be varied. The general focus is placed in the parameter sensitivity analysis which puts more focus on the parameter values. On the other hand, structural sensitivity analysis puts more focus on the effects of the model structure on the possible outcome.

More focus was put on the type of variations and the frequency of variation. In the course of establishing the variations, more focus was put on the Monte Carlo analyses which draw on the probability distributions and the single alternative values. Secondly, the frequency of variation narrowed down to the on-going change which is regarded as part of the stochastic process. The same case is aligned to the model uncertainty which looks at the impact of the change. Key assumptions are directed at the assumptions of how the entire system works and the kind of the decisions that tap into the abstracts of the system behaviour. The structural sensitivity analyses also provided room for the variation of the model while seeking the trade-offs of the choices. This would be aligned to the uncertainty directed at the current state, which determines the extent of mapping the next state. Some of the approaches placed into consideration include the Particle filter approach and the Kalman’s filter.

Part 2: Modifying the discrete section of the model

In this section, the experiment focused on what would happen if a retailer decides to invest in the new distribution centre. The experiment focused on a base case while runs at zero without considering the investment in the advanced technology. The modification narrowed to a few parameters such as selection of the percentages for the deliveries, which came direct from the new distribution centre. Besides, the experiment focused on a similar process which would help in planning the details and consider workers as part of the most paramount factors. Other provisions include the loading requirements, packing requirements that need the support of 1 worker and 3 workers needed to work with both the frozen or chilled goods.

Another concern is that around 20% of the goods in the new centre are in the freezer while half of them will not be defrosted before doing the parking. Around 30% of the goods are in the chilled storage as the rest of the 70% are experience to be served with the most ambient temperature warehouse. The process times are anticipated to be in the base model while defrosting alone would take around 5 hours. The model under consideration focuses on the course of determining the number of workers that need to be employed to handle at least 75% of the goods available at the new distribution centre. The model requires modifications to suit the new operations while trying to convince the manager in charge that the functions would not impeded by any factor or constraint. First, the simulation process focused on how the new distribution centre is expected to look like in terms of the layout and design. The figure below shows the GUI.

It could be noted that for consistent flow of goods in the shopping centre, it is required that the base model provides enough space for the flow of goods in and outside the parking area. The same attention is given to the time taken by the trucks at the new distribution centre, the parking time, offloading and the delay time between the preceding and the succeeding truck. The model encompasses a range of the input parameters that are likely to be retrieved from a range of sources such as estimations, the real observations, the statistical process and expert-based approaches that use analytical tools (Hao and Shen 2008). Java classes were equally necessary. Some of the parameters extracted from the sources include the service time, which also the time that would be spend on a service.

Another parameter is the schedule, which encompasses the working hours, natural holidays and the working days. Nevertheless, more observations are drawn towards the quantity of resources which are served via the service providers in the new distribution centre. The model focused on two entities. The first one focused on the storage parameter. The type of characters incorporated includes the ID, which should be an integer and described as an internal identifier of the storage type. The second one is the in PortTime, which should be a double and demotes the time when the goods are introduced in the freezer or elsewhere. Another parameter is the onBoardCont which includes the tag on goods that off loaded from a particular point in time. The inBlockTime would also denote the parameter used in denoting the delay time identified or approximated for every block. The model in place expresses the storage services and shopping process put in place.

It is worth noting that the model fostered simulation of the incoming processes with the standard blocks of the AnyLogic model put in use. If such issues like the queue block, then the AnyLogic simulation should avail alternatives that can enhance the process. At some point, more focus will be placed on the simulation loading process that focuses on the trucks reporting at the centre and the ones leaving the parking space. Two parameters can be introduced to cater for the adjustments. The parameters include the truck_waiting and the truck_storage.

Part 3: Modifying the system dynamics as part of the model

This section focus on what would happen in the next 10 years if the retailer, who invested in the new distribution centre, chooses to invest around 3% of the financial resources in the IT systems and processes as part of the advanced technology. This takes note of the fact that there is a percentage of the goods that come from the distribution centre model adopted in part 2. It should be noted that the model that adopts the advanced technology considers measurements in terms of the advanced technology units as well as the development costs that should be reasoned around £100000. The estimations are based on the assumption of the continuous variable as well as it is presumed that the technology is mature enough to make it operational in the next 2 years (Mazhari et al. 2011).

The model presumed that the advanced operational technology is likely to be obsolete after working for 12 years. Other assumptions are that the costs of sales would fall by at least 1% for every of the advanced technology unit. The means that with costs standing at 99% of what would have been experienced in every unit of advanced technology, then the subsequent sale would encounter around 98.01% chances of fetching the market. The new technology would be regarded as an additional resource that would change the operations at the new distribution centre with the help of the Anylogic simulation process. Below is the R&D model that takes into consideration the IT resources that would be engaged in automating the significant processes within the centre.

Therefore, in the process of simulation, the Anylogic tool is better placed as the most appropriate tool that will optimize the warehouse layout as well as operation, forecasting and at the same time, adapting to the operational needs. It is worth noting that the simulation modelling may still apply the warehouse simulation software in introducing flexible capabilities while investing in the real world structure, resources and processes at the same time. This would move the system from being open loop to closed loop automated system.

Part 4: Modified state charts that refines the ABM sub module

The experiment puts into consideration the detailed market research regarding the customer behaviour. The design first tapped into brand attraction as well as the effects of the word of mouth which would only convince around 30% of the purchase while the rest would be influenced with research that is expected to run for around 5-10 days. The state chart also considers the fact that only two thirds would decide to make a purchase, then this would imply that half of them are likely to fail making any purchase and the remaining will look upon a convincing future. Around 25% of the adopters are equally making a move that would see them purchase in the future.

Dig deeper into A Report on How Big Data Tools Should Be Used and For with our selection of articles.

References

• Bazan, P. and German, R., 2012, December. Hybrid simulation of renewable energy generation and storage grids. In Proceedings of the 2012 Winter Simulation Conference (WSC) (pp. 1-12). IEEE.
• Fu-gui, D.O.N.G., Hui-mei, L.I.U. and Bing-de, L.U., 2012. Agent-based simulation model of single point inventory system. Systems Engineering Procedia, 4, pp.298-304.
• Faria, F. and Reis, V., 2015, September. An Original Simulation Model to Improve the Order Picking Performance: Case Study of an Automated Warehouse. In International Conference on Computational Logistics (pp. 689-703). Springer, Cham.
• Mazhari, E., Zhao, J., Celik, N., Lee, S., Son, Y.J. and Head, L., 2011. Hybrid simulation and optimization-based design and operation of integrated photovoltaic generation, storage units, and grid. Simulation Modelling Practice and Theory, 19(1), pp.463-481.
• Hao, Q. and Shen, W., 2008. Implementing a hybrid simulation model for a Kanban-based material handling system. Robotics and Computer-Integrated Manufacturing, 24(5), pp.635-646.