JP3940684B2 - Capital investment risk assessment method and system, and program therefor - Google Patents

Description

【０１１５】
以上のような確率的不確実環境のパラメタ生成手法によれば、ユーザが確率的不確実環境の要素や設備投資を考慮する期間を指定するだけで、その最終期における各要素の想定分布から各要素の各変動の変動率を示すパラメタを逆算して自動生成することができる。したがって、ユーザ自身が確率的不確実環境のパラメタを生成する必要なしに、有効なパラメタを効率よく自動生成することができる。
【０１１６】
［２−６．リスク評価基準生成］
本実施形態の設備投資リスク評価システム１００においては、前述したように、決定木の評価にユーザのリスク選好基準を表すリスク評価基準１２２を使用するが、リスク評価基準１２２の設定は容易ではない。そこで、本実施形態においては、リスク評価基準生成手段１０９により、リスク評価基準のどの選択肢を選ぶかをユーザに指定させることにより、ユーザの持つリスク選好度を学習する。
【０１１７】
図２３は、ユーザのリスク選好度を学習するためのリスク評価基準生成手段１０９の詳細な構成を示すブロック図である。この図２３に示すように、リスク評価基準生成手段１０９は、情報提示手段２３０１、意思決定手段２３０２、ユーザ選択結果取得手段２３０３、学習手段２３０４、を備えている。
【０１１８】
この構成により、リスク評価基準生成手段１０９は、決定木評価手順のステップＳ１００４（図１０）において、与えられたリスク評価基準１２２に基づき計算されたリスク調整済価値の大きさで選択肢を選択する代わりに、情報提示手段２３０１で各選択肢のリスクとリターンをユーザに提示し、ユーザ選択結果取得手段２３０３でユーザの選択結果を取得する。ユーザが選択した選択肢に関する情報は、決定結果２３１１として、意思決定手段２３０２により決定木評価手段１０３に通知する。
【０１１９】
このように、ユーザの選択結果に基づく意思決定を行うことにより、ユーザのリスク選好度が間接的かつ部分的に認識できるため、ユーザの選択結果から導出できる部分的なリスク選好度を学習手段２３０４で学習し、評価ルール２３１２として蓄積する。例えば、図２４に示すように、リスク評価基準が関数ｆで表現される場合、ユーザの１つの選択結果からｆに関する制約式が生成できる。具体的には、図２４の選択肢１より選択肢２の価値が高いとユーザが判断することにより、制約式「選択肢１のリスク調整済価値＜選択肢１のリスク調整済価値」が学習できる。この処理を繰り返すことにより、ユーザのリスク選好度を学習した結果としての制約式の集合である評価ルール２３１２からリスク評価基準１２２を生成することができる。
【０１２０】
具体的には、蓄積された制約式の集合から数理計画法および数値解法によりｆの推定を行うことができる。どのタイプの数理計画法を適用できるかは、ｆの関数形に依存するが、制約式を満たす関数系の選択を行い、関数のパラメタを数理計画法で解く。解けない場合は、数値解法で近似的に解くことも可能である。また、ユーザに問い合わせるまでもなく、過去に蓄積された評価ルール２３１２に基づいて選択肢の選択が可能な場合もある。この場合には、ユーザ判断手段２３０３によるユーザの選択結果の取得を省略し、意思決定手段２３０２により評価ルール２３１２に基づいて自動的に意思決定を行うことができる。
【０１２１】
以上のようなリスク評価基準生成手法によれば、ユーザがリスク評価基準の選択肢に関する情報を参照してリスク評価基準を選択するだけで、リスク評価基準を自動生成することができる。したがって、ユーザ自身がリスク評価基準を生成する必要なしに、ユーザのリスク選好度を反映したリスク評価基準を効率よく自動生成することができる。
【０１２２】
［３．他の実施形態］
なお、本発明は、前述した実施形態に限定されるものではなく、本発明の範囲内で他にも多種多様な形態が実施可能である。まず、各図に示したシステム構成、各部の構成、処理手順、データ構造、表示形式等は一例にすぎず、具体的な構成や処理手順、データ構造、表示形式等は自由に選択可能である。さらに、本発明は、前述した通り、鉄鋼、化学、電力プラントなどの比較的長期間の設備投資の意思決定に好適であるが、それに限らず、各種の事業における各種の設備投資の意思決定を行う場合に幅広く適用可能であり、いずれの場合でも、同様に優れた効果が得られるものである。
【０１２３】
【発明の効果】
以上説明したように、本発明によれば、外部環境を表すシナリオを確率的なものと非確率的なものに分けて決定木の生成と評価を行うことにより、多様な外部状況を合理的かつ効率的に設定可能で、しかも、多様な外部状況における評価結果を総合的に考慮しながら設備投資の意思決定を合理的かつ効率的に行うことが可能な設備投資リスク評価方法とシステムを提供することができる。
【図面の簡単な説明】
【図１】本発明を適用した基本的な実施形態に係る設備投資リスク評価システムの全体構成を示すブロック図。
【図２】図１に示す設備投資リスク評価システムによる設備投資リスク評価の処理の概要を示すフローチャート。
【図３】図１に示す設備投資リスク評価システムにより生成される決定木の一例を示すデータ表示図。
【図４】図３に示す決定木の生成に使用されるマクロ外部シナリオの集合の一例を示すデータ表示図。
【図５】図３に示す決定木の生成に使用される確率的不確実環境として、確率的不確実環境の各要素における１期間の変動を表現する二分木の一例を示すデータ表示図。
【図６】図３に示す決定木の生成に使用される確率的不確実環境として、図５に示す確率的不確実環境の各要素の変動を示す二分木を複数期間に亘って組み合わせた場合の一例を示すデータ表示図。
【図７】図３に示す決定木の生成に使用される戦略オプションの一例を示すデータ表示図。
【図８】図２に示す決定木生成手順のサブルーチンを示すフローチャート。
【図９】図２に示す決定木評価で使用するリスク評価基準の一例を示すグラフ。
【図１０】図２に示す決定木評価手順のサブルーチンを示すフローチャート。
【図１１】図１０に示す決定木評価手順中で行うリターン・リスク計算手順のサブルーチンを示すフローチャート。
【図１２】図１０に示す決定木評価手順により得られる評価結果のデータ構造を示すデータ構造図。
【図１３】図１に示す評価結果分析表示手段の詳細な構成を示すブロック図。
【図１４】図２に示す表示結果分析表示手順のサブルーチンを示すフローチャート。
【図１５】図１４に示す表示結果分析表示手順における評価結果集合表示の一例として、各要素を各次元とする３次元表形式の表示例を示すデータ表示図。
【図１６】図１４に示す表示結果分析表示手順において複数の評価結果を指定した場合の個別評価結果表示の一例を示すデータ表示図。
【図１７】図１４に示す表示結果分析表示手順におけるシミュレーション結果表示の一例を示すデータ表示図。
【図１８】図１に示す非確率的不確実環境生成手段と評価結果蓄積手段の詳細な構成を示すブロック図。
【図１９】図１８に示す非確率的不確実環境生成手段によるマクロ外部シナリオの作成例を示す説明図。
【図２０】図１８に示す評価結果蓄積手段による評価結果の検査に使用される許容リスク仕様の一例を示すデータ表示図。
【図２１】図１８に示す評価結果蓄積手段による検査結果をフィードバックして非確率的不確実環境生成を行う場合の確認不要なシナリオ空間と確認の必要なシナリオ空間を示す概念図。
【図２２】図１に示すパラメタ生成手段による確率的不確実環境のパラメタ生成に使用する想定分布情報を示すデータ表示図。
【図２３】図１に示すリスク評価基準生成手段の詳細な構成を示すブロック図。
【図２４】図２３に示すリスク評価基準生成手段による選択肢の提示、ユーザの選択結果に基づく評価ルールの蓄積、の一例を示す説明図。
【符号の説明】
１００…設備投資リスク評価システム
１０１…評価ケース設定手段
１０２…決定木生成手段
１０３…決定木評価手段
１０４…評価結果蓄積手段
１０５…評価結果分析表示手段
１０６…キャッシュフロー計算手段
１０７…非確率的不確実環境生成手段
１０８…パラメタ生成手段
１０９…リスク評価基準生成手段
１１１…非確率的不確実環境（マクロ外部シナリオ）
１１２…確率的不確実環境
１１３…戦略オプション
１２１…決定木
１２２…リスク評価基準
１２３…評価結果
３０１…意思決定ノード
３０２…外部状況指定ノード
３０３…外部確率選択ノード
３０４…シナリオ
３１０…開始ノード
３１１…意思決定分枝
３１２…外部状況指定分枝
３１３…外部確率選択分枝
１２０１…評価ケース
１２０２…結果情報
１２１１…マクロ外部シナリオ
１２１２…戦略オプション
１２１３…リスク評価基準
１２２１…割引現在価値期待値
１２２２…割引現在価値分布
１２２３…意思決定結果情報
１３０１…表示条件設定手段
１３０２…評価結果集合表示手段
１３０３…個別評価結果表示手段
１３０４…シミュレーション手段
１３０５…シミュレーション結果表示手段
１３０６…総合評価決定手段
１８０１…非確率的不確実環境のシナリオ要素
１８０２…基本シナリオ
１８０３…マクロ外部シナリオ生成手段
１８１１…許容リスク仕様
１８１２…検査手段
１８１３…蓄積手段
２３０１…情報提示手段
２３０２…意思決定手段
２３０３…ユーザ選択結果取得手段
２３０４…学習手段
２３１１…決定結果
２３１２…評価ルール[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a capital investment risk evaluation method and system for evaluating a capital investment plan in consideration of risks in a situation where the business environment is uncertain.
[0002]
[Prior art]
Since the bubble burst, capital investment in the manufacturing industry has been declining throughout Japan. This is due to the impact of the adjustment process of excess equipment due to the decline in demand, but it seems that there are many cases where it is shrinking more than necessary for investment in the uncertain future. The high economic growth period was an era when it could be made, and there was little need to judge whether or not to invest while strictly considering uncertainty (risk) regarding capital investment. However, in the situation where capital investment decision-making cannot be made due to the vague anxiety about the future as described above, there is a need for a method that supports a reasonable capital investment plan with sufficient consideration of uncertainty (risk). Yes.
[0003]
For example, Patent Document 1, Non-Patent Document 1 to Non-Patent Document 5 and the like are described with respect to an optimal equipment investment plan (optimal power supply plan) and an optimal equipment operation plan (economic operation plan) in consideration of uncertainties in the power business. As you can see, much research has been done.
[0004]
Non-Patent Document 1 proposes a framework for a long-term power facility planning technique. However, uncertain factors are modeled assuming a certain situation (realized value), and uncertainty is not explicitly considered. Japanese Patent Laid-Open No. 2004-228688 calculates a future expected power demand and derives a countermeasure work plan for a facility that exceeds the suppliable capacity.
[0005]
Non-Patent Document 2 constructs a power supply planning problem assuming several situations for uncertain demand. However, the power supply planning model is unrealistic and complicated, and requires a parallel computer, so it is not at a practical level. Non-Patent Document 3 is building a power plan problem by assuming several situations for uncertain power fixed costs / variable costs, but in order to derive the optimal power configuration in the target year, decision making Is performed only once.
[0006]
Non-Patent Document 4 proposes a statistical power planning method, but this method cannot handle a scenario without a statistical basis. Non-Patent Document 5 describes a non-linear programming method and dynamic planning as an optimal power supply plan formulation program, taking into account network constraints and load unequal time between each system, using an optimal power interchange simulation function in multiple regions (systems). We propose a method that can formulate a minimum cost power development plan. In this approach, information about the uncertain future situation needs to be explicitly provided by the user.
[0007]
As for risk assessment methods in capital investment, decision tree analysis and real option methods have been proposed as described in Non-Patent Documents 6 and 7, Patent Documents 2 and 3, and the like.
[0008]
Non-Patent Document 6 proposes a decision tree creation and evaluation support tool based on spreadsheet software, but in this method, the decision tree is not relatively simple and is difficult to actually use. Non-Patent Document 7 describes a real option method for evaluating the value of a strategic option.
[0009]
Patent Document 2 proposes a system that supports management decision making by reflecting the future management state of a company in the present value. According to this method, the cash flow can be calculated by referring to the cash flow calculation module after inputting the simulation conditions. Patent Document 3 proposes a repayment plan support system and a capital investment plan support system that can make a repayment plan and a capital investment plan so that the repayment of borrowings does not impose management.
[0010]
In the conventional risk evaluation methods as described above, the detailed cash flow model at the time of facility operation is outside the scope of the system, and in many cases, it is assumed that the user gives the cash flow model.
[0011]
[Patent Document 1]
JP 2001-112171 A
[Patent Document 2]
JP 2001-125962 A
[Patent Document 3]
Japanese Patent Laid-Open No. 11-134387
[0012]
[Non-Patent Document 1]
BF Hobbs, "Optimization methods for electric utility resource planning", European Journal of Operational Research 83, 1995, pp. 1-20
[Non-Patent Document 2]
Nishikawa, Tezuka, Kita, Nakano, "Optimal power supply planning under uncertain demand using scenario integration algorithm", IEEJ Transactions, Vol. 112-C, no. 6, 1992, p. 356-363
[0013]
[Non-Patent Document 3]
Tanabe, Yasuda, Yokoyama, “Power Robust Mix Determination Method Considering Uncertainty of Fixed and Variable Power Supplies”, IEEJ Transactions, Vol. 112-B, no. 4, 1992, pp. 331-338
[Non-Patent Document 4]
Supervised by Ryuichi Yokoyama, “Power Liberalization and Technology Development (Chapter 7 Section 4: Power Supply Expansion Planning Method Considering Uncertainty)”, Tokyo Denki University Press, 2001, pp. 203-217
[Non-Patent Document 5]
"ESPRIT: Minimum Cost Optimal Power Plan Planning Program for Interconnected Systems", [online] Power Development Co., Ltd., Development Calculation Center, Inc., [February 20, 2003 search], Internet <http: //www.kcc .co.jp / esprit /. ＞
[0014]
[Non-Patent Document 6]
"Precision Tree", [online], Palisade Corporation, [searched February 20, 2003], Internet <http://www.palisade.com/html/ptree.html >
[Non-Patent Document 7]
L. Trigeorgis, "Real Options", MIT Press, 1996 (Yuichiro Kawaguchi et al., "Real Options", Economist, 2001)
[0015]
[Problems to be solved by the invention]
As described above, in the prior art, capital investment risk assessment methods that take into account the uncertainty when external conditions are given have been proposed, but these methods are applied to actual capital investment risk assessments. However, it is very difficult to rationally set various external situations, and it is very difficult to make a reasonable decision on capital investment while comprehensively considering the evaluation results in various external situations. This will be described below.
[0016]
First, in the above-described prior art, although an uncertain external situation scenario is considered, it is not clear whether the scenario can be handled probabilistically or not. For example, among uncertain external situations in the electric power business, future demand, fluctuations in raw material prices, etc. can be expressed stochastically, but the progress of deregulation regarding power liberalization and the status of new entry into the electric power business Etc. cannot be expressed stochastically.
[0017]
Thus, in order to make an optimal decision on a capital investment plan in consideration of external situations with different characteristics, it is necessary to explicitly handle uncertain external situations that cannot be expressed probabilistically separately from probabilistic ones. There is no such prior art. For this reason, in the actual capital investment risk assessment using the conventional technology, it was difficult to create a rational scenario that appropriately represented various external situations, taking into account their probabilistic and non-probabilistic properties. . In particular, when external situations are complex, the possible external situations increase exponentially, and there are a huge number of scenarios that must be considered.
[0018]
  The present invention has been proposed in order to solve the problems of the prior art as described above.By explicitly handling uncertain external situations that cannot be expressed probabilistically by distinguishing them from probabilistic ones, the probability and non-probabilistic nature of various external situations can be handled without the need for users to generate scenarios. It is possible to rationally and efficiently generate scenarios that are expressed appropriately based onMoreover, it is to provide a capital investment risk evaluation method and system capable of rationally and efficiently making a decision on capital investment while comprehensively considering evaluation results in various external situations.
[0019]
[Means for Solving the Problems]
The present invention can generate various external situations rationally and efficiently by generating and evaluating a decision tree by dividing scenarios representing the external environment into probabilistic and non-probabilistic ones. This makes it possible to rationally and efficiently make capital investment decisions while comprehensively considering evaluation results in external situations.
[0020]
  The invention of claim 1 is a capital investment risk evaluation method for evaluating a capital investment plan in consideration of risks in an uncertain external environment using a computer having an input / output unit, a storage unit, and an arithmetic processing unit. ,A non-stochastic uncertainty environment that represents the uncertainty of a non-stochastic external environment that cannot be statistically modeled as a stochastic process, or a stochastic that represents an uncertainty of a stochastic external environment that can be statistically modeled as a stochastic process When defined as an uncertain environment, a strategy option that represents an option to be evaluated by a user, and a risk evaluation standard that represents a user's risk appetite, the non-realization realized by the arithmetic processing unit of the computer and the program that controls it Non-stochastic uncertain environment generation step, decision tree generation step, decision tree evaluation step, evaluation result as steps performed using the probabilistic uncertain environment generation means, decision tree generation means, decision tree evaluation means, evaluation result storage means It has the accumulation step. Here, in the non-stochastic uncertain environment generation step, a scenario of a non-stochastic uncertain environment is generated by the non-stochastic uncertain environment generation means based on an instruction of a scenario element from the user input by the input / output unit. This is a step of determining a set of elements and generating a scenario of a non-stochastic uncertain environment by combining scenario elements from the set. The decision tree generation step includes: a stochastic uncertainty environment and a strategy option given as input data from the input / output unit, data stored in the storage unit, or data generated from any of the data; and the non-probability This is a step of generating a decision tree by the decision tree generation means using the scenario of the non-stochastic uncertainty environment generated in the static uncertainty environment generation step. The decision tree evaluation step is performed by the decision tree evaluation means based on a risk evaluation criterion given as data generated from input data from the input / output unit, data stored in the storage unit, or any of the data. In this step, the decision tree generated in the decision tree decision step is evaluated. The evaluation result accumulation step is a step in which the evaluation result of the decision tree obtained by the decision tree evaluation step is stored in the storage unit by the evaluation result accumulation unit. Further, the decision tree generation step includes a decision node that generates a decision branch corresponding to each option of the strategy option, and an external situation that generates an external situation designation branch corresponding to each scenario of the non-stochastic uncertainty environment. Generating a decision tree using three types of nodes: a designated node and an external probability selection node that generates an external probability selection branch corresponding to each variation of each element of the stochastic uncertainty environment. The decision tree evaluation step calculates a return and a risk for each decision branch corresponding to each option of the strategy option for the decision tree generated in the decision tree determination step, and calculates the return and risk from the calculated return and risk. Calculating a risk-adjusted value based on risk assessment criteria and selecting a strategy option option that maximizes the risk-adjusted value.
[0021]
  Claim 9The invention of claim 1 grasps the invention of claim 1 from the viewpoint of the system, and uses a computer to evaluate a capital investment plan in consideration of risks in an uncertain external environment. InA non-stochastic uncertain environment generation means, a decision tree generation step, a decision tree generation step, a decision tree evaluation step, a non-stochastic uncertainty environment generation means having a function corresponding to the evaluation result accumulation step, a decision tree generation means; It is characterized by comprising decision tree evaluation means and evaluation result storage means.
[0022]
  Claim 10The invention of claim 1,Claim 9The invention is understood from the viewpoint of a computer program, and a capital investment plan is evaluated in consideration of risks in an uncertain external environment using a computer having an input / output unit, a storage unit, and an arithmetic processing unit. In the program for evaluating capital investment risks,Non-stochastic uncertainty environment generation step, decision tree generation step, decision tree evaluation step, evaluation result accumulation step, non-stochastic uncertainty environment generation function, decision tree generation function, decision tree evaluation function in the invention of claim 1 The evaluation result storage function is realized by a computer.
[0023]
  According to the invention as described above, the generation and evaluation of the decision tree can be performed reasonably and efficiently by dividing the scenario representing the external environment into a probabilistic one and a non-stochastic one. In other words, probabilistic scenarios can be decided based on statistics such as the average and variance of expected gains, but for non-stochastic scenarios, all cases need to be checked. The decision tree can be generated and evaluated reasonably and efficiently. And by generating and evaluating decision trees reasonably and efficiently in this way, a huge number of rational scenarios can be efficiently and automatically generated according to various external situations. Comprehensive evaluation can be performed by calculating the decision and cash flow, and the evaluation result can be analyzed in multiple ways. Accordingly, it is possible to rationally evaluate the strategy option given by the user and to make an optimum decision on capital investment.
  In addition, simply by specifying scenario elements for a non-stochastic uncertain environment, a set of scenario elements can be automatically determined, and a scenario for a non-stochastic uncertain environment can be automatically generated by combining the scenario elements. The scenario of the non-stochastic uncertainty environment reflecting the user's intention can be efficiently and automatically generated without the user himself / herself needing to generate the scenario of the non-stochastic uncertainty environment. In addition, since the user's risk appetite can be reflected in the evaluation of the decision tree, it is possible to make an optimum decision according to the user's risk appetite.
[0024]
  The invention of claim 2 is the capital investment risk evaluation method of claim 1,The evaluation result accumulation step and the non-stochastic uncertainty environment generation step include the following steps. That is, the evaluation result accumulating step determines whether or not the evaluation result of the decision tree by the decision tree evaluation step is redundant and need not be confirmed based on an allowable risk specification set in advance as a condition regarding the redundant and unnecessary confirmation result. An inspection step for inspecting and eliminating an evaluation result that does not require confirmation, and an accumulation step for storing the evaluation result remaining without being excluded by this inspection step are included. Further, the non-stochastic uncertain environment generation step includes a step of generating a non-stochastic uncertain environment scenario by selecting a combination of scenario elements based on the information obtained by the inspection step.
According to the present invention, by setting in advance conditions related to evaluation results that are redundant and needless confirmation, the number of non-stochastic uncertainty environment scenarios becomes enormous, and even when enormous evaluation results are obtained, The number of evaluation results can be reasonably reduced by inspecting the evaluation results and automatically excluding evaluation results that do not require confirmation. Accordingly, the user can confirm only useful evaluation results that do not include redundant and unnecessary confirmation results without having to confirm all evaluation results. Then, by repeating the setting of evaluation units, generation and evaluation of decision trees, and inspection of evaluation results, learning the range of effective scenarios while feeding back the inspection results of evaluation results, and generating more appropriate scenarios Makes optimal investment decisions.
[0025]
  The invention of claim 3Claim 1 or claim 2In the capital investment risk assessment method ofIt is characterized by having an evaluation unit setting step and an evaluation result analysis display step as steps performed using an evaluation unit setting means and an evaluation result analysis display means realized by a calculation processing unit of the computer and a program for controlling it. . Here, in the evaluation unit setting step, based on a user instruction given by the input / output unit, the evaluation unit setting means performs input data from the input / output unit, data stored in the storage unit, or any of them. Each of the three types of elements: a stochastic uncertainty environment and strategy options given as data generated from the data, and a non-stochastic uncertainty environment scenario generated in the non-stochastic uncertainty environment generation step In this step, a combination is set as an evaluation unit and stored in the storage unit. The evaluation result analysis display step generates image data representing an evaluation result obtained by performing the decision tree generation step and the decision tree evaluation step for each evaluation unit by the evaluation result analysis display means, Displaying by the input / output unit;
[0026]
According to the present invention, in accordance with a user instruction, the risk evaluation criterion is combined with two types of elements, that is, a non-stochastic uncertainty environment and a strategy option used for generation of a decision tree, and set as an evaluation unit. The decision tree can be efficiently generated and evaluated for each evaluation unit. Moreover, since the obtained evaluation result can be expressed for each evaluation unit, the user can review the evaluation unit with reference to the evaluation result expressed for each evaluation unit. Therefore, by repeatedly setting the evaluation unit, generating and evaluating the decision tree, and displaying the evaluation result, the user can learn the user's intention and risk appetite while feeding back the evaluation result, and make the optimum decision on capital investment. It can be carried out.
[0027]
  The invention of claim 4Any one of claims 1 to 3In the capital investment risk assessment method ofIt has a risk evaluation standard generation step performed using a risk evaluation standard generation means realized by an arithmetic processing unit of the computer and a program for controlling it. Here, the risk evaluation criterion generating step is a step of presenting information on each option of the risk evaluation criterion to the user through the input / output unit by the risk evaluation criterion generating means, and generating the risk evaluation criterion based on the user's instruction It is.
[0028]
According to the present invention, the risk evaluation standard can be automatically generated only by the user selecting the risk evaluation standard with reference to the information on the options of the risk evaluation standard. Therefore, it is possible to efficiently and automatically generate a risk evaluation standard that reflects the user's risk appetite without the need for the user to generate a risk evaluation standard.
[0033]
  Claim 5The invention ofAny one of Claim 1 thru | or 4In the capital investment risk assessment method ofIt has a parameter generation step performed using parameter generation means realized by an arithmetic processing unit of the computer and a program for controlling it. Here, the parameter generation step is performed by the parameter generation means.When considering the variation of each element of the stochastic uncertainty environment over T (T is a natural number of 1 or more) period, the variation rate of each variation of each element is calculated from the assumed distribution of each element in the T period. This is a step for generating the parameter shown by back calculation.
[0034]
According to this invention, the user can specify a parameter indicating the variation rate of each variation of each element from the assumed distribution of each element in the final period only by specifying a period in which the factors and capital investment of the stochastic uncertainty environment are considered. It can be automatically generated by reverse calculation. Therefore, effective parameters can be automatically generated efficiently without the need for the user to generate parameters for the stochastic uncertainty environment.
[0035]
  Claim 6The invention ofClaims 3 to 5In any of the above-mentioned capital investment risk evaluation methods, the evaluation result analysis display step is characterized. That is, the evaluation result analysis display step includesImage data in a predetermined format expressing the set of evaluation results stored in the storage unit as viewpoints of three types of elements, the non-stochastic uncertainty environment, the strategy option, and the risk evaluation criteria. Generating and displaying by the input / output unit.
[0036]
  Claim 7The invention ofClaim 6In the capital investment risk evaluation method, the predetermined format is a three-dimensional format having three types of elements as dimensions.
[0037]
According to the invention as described above, a set of evaluation results obtained from each evaluation unit that is a combination of three types of elements is converted into three types of elements such as a three-dimensional format having three types of elements as dimensions. Therefore, the user can easily grasp a set of evaluation results from the viewpoint of each element.
[0038]
  Claim 8The invention ofClaim 6 or Claim 7In the capital investment risk evaluation method, the evaluation result analysis display step includes a set display step, an individual display step, and a simulation result display step. here,The set display step is a step of generating image data of the predetermined format expressing the set of the evaluation results and displaying it by the input / output unit. The individual display step generates image data individually representing details of the evaluation result selected from the set of evaluation results based on a user instruction given by the input / output unit, and outputs the input / output This is a step of displaying by the section. Further, in the simulation result display step, based on a user instruction given by the input / output unit, image data expressing a simulation result by performing a simulation using an evaluation result selected from the set of evaluation results is displayed. It is a step of generating and displaying by the input / output unit.
[0039]
According to the present invention, the user can display details of each evaluation result individually or display a simulation result for not only a set of evaluation results but also an evaluation result selected from the evaluation results. It becomes possible to conduct detailed analysis and examination at the individual evaluation result level. Therefore, it is possible to make an optimum decision on capital investment more strictly.
[0040]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be specifically described below with reference to the drawings. However, the embodiment described here does not limit the present invention at all, and merely illustrates one aspect of the present invention.
[0041]
The present invention is typically realized by controlling a computer with software. The software in this case realizes the operational effects of the present invention by physically utilizing computer hardware, and a suitable conventional technique is applied to a portion to which the conventional technique can be applied. Furthermore, specific types and configurations of hardware and software that realize the present invention, a range processed by software, and the like can be freely changed. For example, a program that implements the present invention is one aspect of the present invention.
[0042]
[1. Basic embodiment]
[1-1. Overview of overall configuration and processing]
FIG. 1 is a block diagram showing the overall configuration of a capital investment risk evaluation system according to a basic embodiment to which the present invention is applied.
[0043]
As shown in FIG. 1, the capital investment risk evaluation system 100 first has an evaluation case setting means 101 for setting an evaluation case (evaluation unit), a decision tree generating means 102 for generating a decision tree, and a decision for evaluating a decision tree. A tree evaluation unit 103, an evaluation result storage unit 104 for storing the evaluation result of the decision tree, and an evaluation result analysis display unit 105 for analyzing and displaying the evaluation result are provided. These means 101 to 105 are combined with a cash flow calculation means 106, a non-stochastic uncertainty environment generation means 107, a parameter generation means 108, a risk evaluation reference generation means 109, and the like. The outline of the function of each means 101-109 is as follows.
[0044]
The decision tree generation unit 102 evaluates a non-stochastic uncertain environment 111 representing uncertain external environment uncertainty, a probabilistic uncertain environment 112 representing probabilistic external environment uncertainty, and a user's evaluation. The decision tree 121 is generated using the three types of elements of the strategy option 113.
[0045]
The decision tree evaluation unit 103 evaluates the generated decision tree 121 based on a risk evaluation criterion 122 representing a user's risk preference criterion, and accumulates the evaluation result 123 in the evaluation result accumulation unit 104. When the decision tree is evaluated by the decision tree evaluation unit 103, the cash flow calculation unit 106 calculates the cash flow.
[0046]
The evaluation result analysis display unit 105 is a unit that analyzes the evaluation result 123 stored in the evaluation result storage unit 104 from various viewpoints, and generates and displays image data representing the analysis result. Is presented to the user in an easy-to-understand form to support the user's decision on capital investment.
[0047]
The evaluation case setting unit 101 is a unit that sets an evaluation case (evaluation unit) for generating and evaluating the decision tree 121. That is, the evaluation case setting means 101 is based on a user instruction, and includes three types of risk evaluation criteria 122 used for evaluation of the decision tree 121, non-stochastic uncertainty environment 111 used for generation of the decision tree 121, and strategy option 113. Each combination of elements is set as an evaluation case.
[0048]
The non-stochastic uncertain environment 111 and the stochastic uncertain environment 112 used to generate the decision tree 121 can be created by the user himself / herself or input separately prepared data. Based on the capital investment risk evaluation system 100, it can be created. That is, the non-stochastic uncertain environment generation unit 107 generates the non-stochastic uncertain environment 111 based on the user's instruction, and the parameter generation unit 108 generates the probabilistic uncertain environment based on the user's instruction. A parameter characterizing 112 is generated.
[0049]
The risk evaluation criteria 122 used for evaluating the decision tree 121 can be created by the user himself / herself or input separately prepared data, but is created by the capital investment risk evaluation system 100 based on the user's instructions. You can also That is, the risk evaluation standard generation unit 109 generates the risk evaluation standard 122 based on a user instruction.
[0050]
The capital investment risk evaluation system 100 shown in FIG. 1 stores hardware resources such as an arithmetic processing unit such as a CPU and a storage unit such as a main memory and various memories, which are basically included in the computer, on the main memory. This is realized by controlling with the processed program. In addition, input of instructions from the user and display of data to the user are realized by an input / output unit such as a keyboard, a mouse, and a display that are basically provided in the computer.
[0051]
FIG. 2 is a flowchart showing an outline of processing of capital investment risk evaluation by the capital investment risk evaluation system 100 of the present embodiment. As shown in FIG. 2, based on the user's instruction given by the input / output unit, the evaluation case setting means 101 causes the risk evaluation standard 122, the strategy option 113, the macro external scenario (non-stochastic uncertainty environment 111), Each combination of the three types of elements is set as an “evaluation case (evaluation unit)” (S201). Here, the “macro external scenario” is one expression method for expressing the non-stochastic uncertain environment 111, and details thereof will be described later.
[0052]
The decision tree generation means 102 generates a decision tree 121 using the three types of elements of the stochastic uncertainty environment 112, the strategy option 113 given in the evaluation case, and the macro external scenario 111 (S202). The generated decision tree 121 is evaluated by the decision tree evaluation unit 103 based on the risk evaluation criteria 122 given in the evaluation case, and the evaluation result 123 is stored in the evaluation result storage unit 104 (S203).
[0053]
When the user instructs setting of another evaluation case (YES in S204), the process returns to S201. When the user instructs that there is no other evaluation case (NO in S204), the set of evaluation results 123 stored in the evaluation result storage unit 104 is analyzed by the evaluation result analysis display unit 105, and the analysis result Is displayed (S205). In this case, image data representing an evaluation result obtained by performing decision tree generation and decision tree evaluation for each evaluation case is generated and displayed by an input / output unit such as a display. If the user instructs to review the evaluation case or set another evaluation case (YES in S206), the process returns to S201, and if not (NO in S206), the process ends.
[0054]
[1-2. Basic effects]
According to the capital investment risk evaluation system 100 of the present embodiment as described above, a scenario representing an external environment is divided into a probabilistic and a non-probabilistic one, and a decision tree is generated and evaluated reasonably efficiently. be able to. In other words, probabilistic scenarios can be decided based on statistics such as the average and variance of expected gains, but for non-stochastic scenarios, all cases need to be checked. The decision tree can be generated and evaluated reasonably and efficiently.
[0055]
And by generating and evaluating decision trees reasonably and efficiently in this way, a huge number of rational scenarios can be efficiently and automatically generated according to various external situations. Comprehensive evaluation can be performed by calculating the decision and cash flow, and the evaluation result can be analyzed in multiple ways. Accordingly, it is possible to rationally evaluate the strategy option given by the user and to make an optimum decision on capital investment. In addition, since the user's risk appetite can be reflected in the evaluation of the decision tree, it is possible to make an optimum decision according to the user's risk appetite.
[0056]
Furthermore, according to the user's instructions, each evaluation case can be set as an evaluation case by combining the two types of elements, the non-stochastic uncertainty environment and strategy option used to generate the decision tree, and the risk evaluation criteria. The decision tree can be generated and evaluated efficiently. Moreover, since the obtained evaluation result can be expressed for each evaluation case, the user can review the evaluation unit with reference to the evaluation result expressed for each evaluation case. Therefore, by repeatedly setting the evaluation case, generating and evaluating the decision tree, and displaying the evaluation result, the user can learn the user's intention and risk appetite while feeding back the evaluation result, and make the optimum decision on capital investment. It can be carried out.
[0057]
[2. Details of characteristic processing]
Below, the detail of the characteristic process in each part in the capital investment risk evaluation system 100 of this embodiment is sequentially demonstrated including the definition of each element to handle. In other words, “determination tree generation”, “determination tree evaluation”, and “evaluation result analysis display” in the outline of the processing shown in FIG. 2 will be described in detail, and elements used for generation and evaluation of decision trees are generated. “Generation of non-stochastic uncertain environment”, “Parameter generation of probabilistic uncertain environment”, and “Generation of risk evaluation criteria” will be described in detail.
[0058]
[2-1. Decision tree generation]
The capital investment plan targeted by the present invention is a relatively long-term plan for steel, chemicals, power plants, etc., and makes flexible decision-making at each point in response to future uncertain changes in the external environment. It is assumed that this is done. In other words, the “plan” in the present invention is not a sequence plan that “installs equipment A in XX year and updates equipment B in △△ year”. If so, the plan is to include an option based on an uncertain external environment. The decision tree generated in the present embodiment is one expression format for expressing such a plan. FIG. 3 shows an example of the decision tree 121.
[0059]
The decision tree 121 shown in FIG. 3 includes three types of branch nodes 301 to 303, which are a decision decision node 301, an external situation designation node 302, and an external probability selection node 303. The decision making node 301 is a branch that can be independently selected by the decision maker from the given strategy options 113. Further, the external situation designation node 302 and the external probability selection node 303 are branch nodes that are selected depending on the external environment, and a non-stochastic external situation designation node depending on whether or not the selection is performed stochastically. 302 and a probabilistic external probability selection node 303. The result of optimal decision making in a given external environment can be expressed as a decision result scenario 304 by a path on the decision tree 121.
[0060]
As described above, such a decision tree 121 is obtained by the decision tree generation unit 102 by using the non-stochastic uncertainty environment 111 representing the uncertainty of the non-stochastic external environment and the uncertainty of the probabilistic external environment. It is generated using three types of elements: a stochastic uncertainty environment 112 to represent, and a strategy option 113 that the user wants to evaluate. In this specification, these elements are defined as follows.
[0061]
“Non-stochastic uncertainty environment” means an uncertain external situation that cannot be controlled by the decision maker and that cannot be statistically modeled as a stochastic process. For example, the progress of deregulation related to the liberalization of electricity in the electric power business and the status of new entry into the electric power business are non-stochastic environments. It is possible to assign an occurrence probability based on the subjectivity to such a non-stochastic uncertainty environment, but the "probability" in that case does not have a statistical basis, so it is statistically modeled. It should be clearly distinguished from possible external situations.
[0062]
There are several ways of expressing this “non-stochastic uncertainty environment”, but in this embodiment, as an example, it is expressed as a set of “macro external scenarios (sequences of events)” as shown in FIG. is doing. In FIG. 4, three macro external scenarios are listed as examples of macro external scenarios regarding the progress of deregulation related to power liberalization.
[0063]
“Probabilistic uncertain environment” means an uncertain external situation that cannot be controlled by a decision maker and that can be statistically modeled as a stochastic process. For example, future demand in the power business and fluctuations in raw material prices are probabilistic uncertain environments. There are several ways of expressing this “stochastic uncertainty environment”, but in this embodiment, a binary tree expression as shown in FIG. 5 is used as an example.
[0064]
The binary tree shown in FIG. 5 represents demand and raw material price fluctuations for one period as elements of the stochastic uncertainty environment. The variation characteristics of each element can be expressed by “parameter (variation rate and occurrence probability of each variation)”. In FIG. 5, there are two demand parameters: “variation rate: 3% increase, occurrence probability: 0.5”, “variation rate: 1% decrease, occurrence probability: 0.5”, and the raw material There are two price parameters: “change rate: 2% increase, occurrence probability: 0.5” and “change rate: 2% decrease, occurrence probability: 0.5”.
[0065]
When the fluctuation characteristics do not change over a plurality of periods to be considered, the fluctuations of the plurality of periods can be easily expressed as shown in FIG. 6 by simply combining the fluctuations of one period. This FIG. 6 has shown an example at the time of combining the binary tree of the demand and raw material price shown in FIG. 5 over several periods.
[0066]
A “strategy option” is a set of options that a decision maker can control. For example, the establishment of a thermal power plant or the implementation of large-scale repairs in the electric power business are strategic options. Although it is possible to set a set of all options that are possible in principle as a strategy option, it takes cost and preparation to prepare each option, and in reality it is a set of all options that are possible in principle. The included subset is a strategy option. FIG. 7 shows an example of a strategy option. In FIG. 7, three strategic options are listed as an example of a strategic option regarding whether or not a power plant is newly established, discarded, or repaired in the electric power business.
[0067]
FIG. 8 is a flowchart showing a generation procedure of the decision tree 121 by the decision tree generation means 102, and corresponds to a subroutine of the decision tree generation procedure (S202) shown in FIG. Note that FIG. 8 illustrates a case in which decision tree branches and nodes are generated based on the order of “width-first search”, but this procedure is merely an example, and other processes such as “depth-first search” and the like. It is also possible to generate a decision tree in this order. Hereinafter, a procedure for generating the decision tree 121 as shown in FIG. 3 will be described with reference to FIG. The generation of the decision tree 121 is specifically performed by registering the set nodes in the storage unit to form a node database and processing all the nodes registered in the node database.
[0068]
As shown in FIG. 8, in the decision tree generation procedure, first, the start node 310 that is the starting point of the decision tree 121 is set and registered in the node database (S801). The given start node 310 or an unprocessed node is used as a decision node 301, and decision decision branches 311 and nodes corresponding to all options of the strategy options that can be executed in this state are generated (S802). Each node generated corresponding to the decision-making branch 311 from the decision-making node 301 is set as an external situation designation node 302, and an external situation designation branch 312 and a node that can be executed in this state are created (S803). In this case, the external situation designation branch 312 corresponds to a macro external scenario.
[0069]
Each node generated corresponding to the external situation designation branch 312 from the external situation designation node 302 is defined as an external probability selection node 303, and an external probability selection branch 313 and a node that can be executed in this state are generated and generated. Each node is registered in the node database (S804). If the registered node is a node in the period (t period) before the plan end period (T period), it is determined as an unprocessed node in t = t + 1 period. When the registered node is the plan end period (T period), it is not regarded as an unprocessed node. In this case, the external probability selection branch 313 corresponds to a probabilistic uncertain environment, and is given a variation rate and occurrence probability as shown in FIGS.
[0070]
Next, an unprocessed node is selected from the node database (S805). If there is an unprocessed node in the node database (YES in S806), the period is advanced by one period and the process returns to S802. If there is no unprocessed node in the node database (NO in S806), the process ends.
[0071]
Through the above steps S801 to S806, a decision tree 121 as shown in FIG. 3 can be generated. Since a huge number of scenarios are handled, branches of the external probability selection node 303 corresponding to the probabilistic uncertainty environment are comprehensively generated, but branches of the external situation designation node 302 corresponding to the macro external scenario are In the course of the evaluation cycle, it is gradually expanded by the designation of the user who is the evaluator.
[0072]
[2-2. Decision tree evaluation]
Evaluation of the generated decision tree 121 is performed based on the risk evaluation criteria 122 by the decision tree evaluation means 103. Specifically, the “discounted present value expected value”, “discounted present value distribution”, and “decision result information” of the project are calculated and stored in the evaluation result accumulating means 3.
[0073]
In the present specification, the “risk evaluation standard” means an evaluation standard for decision making in selecting an option of strategy options necessary for decision tree evaluation. In the simple case, an evaluation criterion of selecting the option with the highest expected value may be considered, but when considering uncertainty, it is necessary to incorporate not only the expected value but also the expected value variation (variance) into the evaluation. .
[0074]
FIG. 9 is a graph showing functions f1, f2, and f3 for calculating the risk premium ρ from the risk σ as an example of the risk evaluation standard 122. In this example, the risk premium increases exponentially as the risk increases, but the rate of increase differs for each of the functions f1, f2, and f3. For example, it can be considered that the function f1 having a large increase rate is a risk evaluation standard that does not desire risk (low risk appetite) compared to the function f2. The expression method of the risk evaluation standard 122 is not limited to a function, and can be expressed in a table format or the like.
[0075]
FIG. 10 is a flowchart showing an evaluation procedure of the decision tree 121 by the decision tree evaluation means 103, and corresponds to a subroutine of the decision tree evaluation procedure (S203) shown in FIG. FIG. 11 is a flowchart showing a “return / risk calculation procedure” subroutine performed in the procedure of FIG. Here, a procedure for evaluating the decision tree 121 expressing a capital investment project in the T period will be described. In this procedure, each node is evaluated while retreating from the end node (T period) of the decision tree, and finally the start node (1 period) is evaluated. In this case, the value at the start node becomes the “discounted current value expected value” of the current project.
[0076]
Hereinafter, the evaluation procedure of the decision tree 121 as shown in FIG. 3 will be described with reference to FIGS. Specifically, the decision tree 121 is evaluated by registering the choice selected for each decision-making node in each period in the storage unit as a decision-making result and creating a decision-making result database.
[0077]
[2-2-1. Decision tree evaluation procedure]
As shown in FIG. 10, in the decision tree evaluation procedure, first, the value of the variable t indicating the evaluation target period in the T period is set to t = T (S1001). The return and risk of value from the t period to the T period are calculated for each option (decision decision branch 311) of the strategy option in the decision node 301 in the t period (S1002: return risk calculation procedure). Details of this “return risk calculation procedure” will be described later.
[0078]
The risk adjusted value is calculated based on the risk evaluation criteria 122 from the return and risk calculated for each decision branch 311 (S1003: risk adjusted value calculation procedure). Details of this “risk-adjusted value calculation procedure” will be described later. Next, the decision branch 311 that maximizes the risk-adjusted value is selected and accumulated in the decision-result database as the decision-making result, and the risk-adjusted value is set as the current value at the beginning of t (S1004).
[0079]
A series of steps (S1002 to S1004) for evaluating the decision making node in the t period as described above are repeated until t = 1 while subtracting the value of t one by one (NO in S1005, S1006). Then, when t = 1, that is, when the evaluation of the decision-making node in one period is finished (YES in S1005), the process is finished.
[0080]
[2-2-2. Return risk calculation procedure]
As shown in FIG. 11, in the return risk calculation procedure, first, the cash flow in the t period when the external situation designation branch 312 and the external probability selection branch 313 in the t period are given is the combination of each branch. Calculate (S1101). Specifically, the cash flow calculation means 106 can represent the operating status of the equipment and various external environments as constraints, and calculate the cash flow at the time of optimal operation by linear programming.
[0081]
Next, the present value at the beginning of the t + 1 period when the external situation designation branch 312 and the external probability selection branch 313 in the t period are given is obtained only for the combination of each branch (S1102). Specifically, the present value at the beginning of the t + 1 period calculated in step S1004 of the decision tree evaluation procedure for the t + 1 period is acquired. Subsequently, the return and risk of value from the t period to the T period are calculated (S1103). Specifically, the expected value of the amount obtained by adding the cash flow of the t period to the amount obtained by subtracting the present value at the beginning of the t + 1 period obtained for each combination of branches for one period is used as the return, and the variance is the risk.
[0082]
[2-2-3. Risk-adjusted value calculation procedure]
In decision tree evaluation, it is necessary to make strategic option decisions at each time point. In this case, the simplest decision-making criterion is the expected value of the discounted present value obtained by converting the cash flow obtained in the future into the value at the time of decision-making, and it is sufficient to make a decision that maximizes the expected value. However, the expected value does not take into account the variation in cash flow obtained in the future, that is, risk.
[0083]
Therefore, in the decision tree evaluation procedure of this embodiment, the risk adjusted value obtained by incorporating the risk into the expected discounted current value is calculated as the risk adjusted value calculation procedure. That is, the risk-adjusted value is calculated based on the risk evaluation criteria 122 from the return and risk calculated for each decision branch 311. For example, it is assumed that the function f1 is specified in the evaluation case among the risk evaluation criteria shown in FIG.
[0084]
Here, when the risk premium is expressed by a discount rate, the risk-adjusted value can be expressed by the following equation.
[Expression 1]
Risk-adjusted value = Expected discounted value / (1 + Risk premium)
Risk premium = f1 (risk)
[0085]
If the risk premium is expressed as a penalty value, the risk-adjusted value can be expressed by the following formula.
[Expression 2]
Risk-adjusted value = Expected discounted current value-Risk premium
Risk premium = f1 (risk)
[0086]
The decision tree 121 as shown in FIG. 3 can be evaluated by the decision tree evaluation procedure (S1001 to S1006) including the return / risk calculation procedure and the risk-adjusted value calculation procedure.
[0087]
Further, as shown in FIG. 12, the evaluation result 123 obtained by the decision tree evaluation procedure is stored in the evaluation result storage means 104 with a data structure including an evaluation case 1201 and result information 1202. Here, the evaluation case 1201 includes a macro external scenario 1211, a strategy option 1212, and a risk evaluation standard 1213. The result information 1202 includes a discount present value expected value 1221, a discount present value distribution 1222, and a decision result information 1223. Become. Among these pieces of information, especially the decision result information 1223 has a large amount of information, and an enormous storage capacity is required for simple storage. Therefore, it may be stored compactly using a binary decision diagram (BDD).
[0088]
[2-3. Evaluation result analysis display]
The set of evaluation results 123 stored in the evaluation result storage unit 104 is analyzed and displayed by the evaluation result analysis display unit 105. FIG. 13 is a block diagram showing a detailed configuration of the evaluation result analysis display means 105. FIG. 14 is a flowchart showing a display result analysis display procedure by the evaluation result analysis display means 105, and corresponds to a subroutine of the display result analysis display procedure shown in FIG.
[0089]
As shown in FIG. 13, the evaluation result analysis display unit 105 includes a display condition setting unit 1301, a display result set display unit 1302, an individual evaluation result display unit 1303, a simulation unit 1304, a simulation result display unit 1305, and a comprehensive evaluation determination unit 1306. It is equipped with. Hereinafter, the evaluation result analysis display procedure by the evaluation result analysis display means 105 will be described with reference to FIG.
[0090]
In the evaluation result analysis display procedure, first, based on a user instruction given by the input / output unit, display condition setting for narrowing down the display target is performed by the display condition setting means 1301 (S1401). For example, if it is complicated to display all the sets of evaluation results 123 stored in the evaluation result storage unit 104, conditions can be set by the display condition setting unit 1301, and only the evaluation results satisfying the conditions can be displayed. Sensitivity analysis (extraction when the result differs greatly due to the difference of one item) can also be specified as a condition.
[0091]
Next, the display result set display means 1302 displays a set of evaluation results that satisfy the specified display condition (S1402). That is, the set of evaluation results 123 accumulated in the evaluation result accumulating means 104 is displayed in a predetermined format that expresses three elements of the evaluation case (macro external scenario 1211, strategy option 1212, risk evaluation standard 1213) as respective viewpoints. .
[0092]
FIG. 15 is a data display diagram showing a display example of a three-dimensional table format in which each element is each dimension as an example of such an evaluation result set display. As for the display of the set of display results, it is sufficient that at least each of the three elements of the evaluation case is a “viewpoint”, and it is not always necessary to use the table format as shown in FIG. It need not be in dimensional form. However, the display in the three-dimensional table format shown in FIG. 15 is a typical display example that is easy to understand visually.
[0093]
When the user designates one or more of the evaluation results 123 displayed by the evaluation result set display as shown in FIG. 15 (YES in S1403), the individual evaluation result display means 1303 designates the designated evaluation result. The individual evaluation result is displayed individually showing the details (S1404). When multiple evaluation results are specified, comparison is possible. FIG. 16 is a data display diagram showing an example of an individual evaluation result display when a plurality of evaluation results are designated.
[0094]
In addition, when the user instructs a simulation when the user wants to analyze the evaluation result displayed by the individual evaluation result display as shown in FIG. 16 in more detail (YES in S1405), the simulation result is obtained by the simulation unit 1304. A scenario of the investment activity is simulated using the decision-making result information 1223 in 123 and displayed on the simulation result display means 1305 (S1406).
[0095]
FIG. 17 is a data display diagram illustrating an example of the simulation result display. As shown in FIG. 17, in the typical simulation result display, various management indicators are displayed in addition to the macro external scenario, the probabilistic uncertainty environment, and the decision-making result in each period in the specified scenario. To do.
[0096]
When the user referring to the evaluation result set display, the individual evaluation result display, the simulation result display, etc. instructs to review the evaluation case or set another evaluation case (YES in S1407), the process returns to S201 in FIG. If not (NO in S1407), comprehensive evaluation is performed by setting the weight of each strategy option by the comprehensive evaluation determining means 1306 (S1408). In this case, the overall evaluation value can also be calculated by weighted average of discounted present value expectation values by weighting macro external scenarios and risk evaluation criteria which are elements other than the strategy option.
[0097]
According to the evaluation result analysis display method as described above, it is easy to visually understand a set of evaluation results obtained from each evaluation unit, which is a combination of three types of elements, with three types of elements as dimensions. Since it can be displayed in a dimension table format, the user can easily grasp the set of evaluation results from the viewpoint of each element.
[0098]
In addition to the set of evaluation results, the details of each evaluation result can be displayed individually or the simulation result can be displayed for the evaluation result selected from the set. Detailed analysis and examination can be performed at the result level. Therefore, it is possible to make an optimum decision on capital investment more strictly.
[0099]
For example, in the individual evaluation result display shown in FIG. 16, when there is only a strategic option for newly introducing a large-scale plant (simple investment strategy), a strategic option for sequentially increasing a plurality of small plants according to demand (distributed investment) (Strategy) can be compared. In this case, a large facility has a lower cost per unit production and a higher efficiency than a small facility.
[0100]
In the individual evaluation result display shown in FIG. 16, the discount present value distributions of the two strategy options are also displayed as one item of the individual evaluation result display. In this example, the expected cash flow / standard deviation in the diversified investment strategy is ¥ 15.3 billion / ¥ 43.8 billion, the simple investment strategy is ¥ 122.5 billion / ¥ 45.5 billion, and the diversified investment strategy has a higher expected cash flow, In addition, the variance (risk) against changes in the situation is small. Therefore, it can be confirmed that the diversified investment strategy is more effective in an uncertain business environment.
[0101]
[2-4. Generation of non-stochastic uncertainty environment]
In the capital investment risk evaluation system 100 of the present embodiment, as described above, the evaluation case setting unit 101 uses the three types of elements of the risk evaluation standard 122, the strategy option 113, and the macro external scenario 111 based on the user's instruction. Each of these combinations is set as an evaluation case (S201 in FIG. 2), but among these, setting a macro external scenario is not easy. Therefore, in the present embodiment, the non-stochastic uncertain environment generation means 107 supports the creation of a macro external scenario by the user.
[0102]
FIG. 18 is a block diagram showing a detailed configuration of the non-stochastic uncertain environment generation unit 107 and the evaluation result storage unit 104 for supporting creation of a macro external scenario. As shown in FIG. 18, the non-stochastic uncertain environment generation means 107 uses a non-stochastic uncertainty environment scenario element 1801 and a basic scenario 1802 to automatically generate a macro external scenario generation means 1803. It has. Further, the evaluation result accumulating unit 104 is not necessary to confirm with the inspection unit 1812 that performs the inspection on whether the evaluation result 123 obtained by the decision tree evaluation unit 103 is redundant and does not require confirmation based on the allowable risk specification 1811. Accumulating means 1813 for accumulating only the evaluation result 123 is provided.
[0103]
According to the non-stochastic uncertain environment generation unit 107 having the configuration of FIG. 18, the non-stochastic uncertain environment generation unit 107 determines a set of scenario elements 1802 of the non-stochastic uncertain environment based on a user instruction. As a result, the macro external scenario generation unit 1803 can automatically generate the macro external scenario 111 as a combination of the scenario elements 1802.
[0104]
For example, in (1) of FIG. 19, from the three scenario elements 1801, the permutations (_ThreeP_Three) Indicates that six macro external scenarios can be created. If there is a time constraint between the scenario elements 1801, a constraint condition can be described using a Petri net, and a set of macro external scenarios can be expressed as a covering graph.
[0105]
Further, the non-stochastic uncertain environment generation unit 107 prepares a basic scenario 1802 in addition to the scenario element 1801, and the macro external scenario generation unit 1803 adds a scenario element 1802 with the basic scenario 1802 as a core. The macro external scenario 111 derived from the basic scenario 1802 can also be automatically generated. For example, (2) of FIG. 19 shows an example in which one derived scenario is generated from the basic scenario.
[0106]
As described above, the macro external scenario 111 can be automatically generated by the non-stochastic uncertain environment generation unit 107 of FIG. 18, but in reality, the number of combinations becomes enormous, which may cause trouble in the evaluation result analysis display. There is sex. Accordingly, the inspection unit 1812 of the evaluation result accumulating unit 104 shown in FIG. 18 checks whether or not the evaluation result 123 is redundant and does not require confirmation based on the allowable risk specification 1811, and the evaluation result 123 that is redundant and does not require confirmation. And only the evaluation result 123 that does not require confirmation is stored in the storage unit 1813.
[0107]
The allowable risk specification 1811 is set in advance by the evaluation result accumulating unit 104 based on a user instruction. FIG. 20 shows an example of a description of the allowable risk specification 1811, in which an expected value of discounted current value and a range of variance are designated and a comparison with past evaluation results is performed.
[0108]
Further, by feeding back the inspection result of the inspection unit 1812 in the evaluation result accumulating unit 104 to the non-stochastic uncertain environment generation unit 107, the non-stochastic uncertain environment generation unit 107 automatically generates based on the inspection result. By selecting a combination of scenario elements 1801 to be improved, overall processing efficiency can be improved. Specifically, a case may be considered in which a derived scenario related to a scenario element whose result changes greatly is given priority over a scenario element whose result does not change significantly. FIG. 21 is a conceptual diagram showing a scenario space that does not need to be confirmed and a scenario space that needs to be confirmed in the case where non-stochastic uncertainty environment generation is performed by feeding back the inspection result by the evaluation result accumulating means 104.
[0109]
According to the non-stochastic uncertain environment generation method as described above, the user can automatically determine a set of scenario elements 1801 only by inputting or specifying the scenario elements 1801, and a macro external scenario by combining the scenario elements 1801. 111 can be automatically generated. Therefore, the macro external scenario 111 reflecting the user's intention can be efficiently and automatically generated without the user himself / herself generating the macro external scenario 111.
[0110]
In addition, when the user sets the permissible risk specification 1811 that is a condition related to the evaluation result that is redundant and does not require confirmation, the number of the macro external scenarios 111 becomes enormous and the enormous evaluation result 123 is obtained. However, the evaluation results 123 can be inspected, and the evaluation results that do not require confirmation can be automatically excluded to rationally reduce the number of evaluation results.
[0111]
Therefore, the user can confirm only useful evaluation results that do not include redundant and unnecessary confirmation results without having to confirm all the evaluation results, and the scenario element 1801 or the like based on the inspection results of the evaluation results. The set can be reviewed and reset. Therefore, in the capital investment risk evaluation system 100 according to the present embodiment, by repeating the setting of the evaluation unit, the generation and evaluation of the decision tree, and the inspection of the evaluation result, an effective scenario can be obtained while feeding back the inspection result of the evaluation result. Learn the scope and generate more appropriate scenarios to make optimal investment decisions.
[0112]
[2-5. Parameter generation of stochastic uncertainty environment]
In the capital investment risk evaluation system 100 of the present embodiment, as described above, the stochastic uncertainty environment 112 is used to generate the decision tree, but it is not easy to set parameters of the stochastic uncertainty environment 112. Therefore, in this embodiment, the parameter generation unit 108 reversely generates parameters from the assumed distribution of each element in the final period.
[0113]
Specifically, a decision tree parameter can be generated by giving assumed distribution information (a distribution name such as a normal distribution, an average value and a variance value which are distribution parameters defining the distribution). For example, as shown in FIG. 22, assuming a normal distribution as the demand distribution in the final period (T period) of the capital investment project, the estimated average value μ and variance σ of the future demand amount²Is specified. Then, by solving the polynomial equation system shown in the following equations 1 and 2 using a numerical solution method such as Newton's method, the demand fluctuation rate (increase rate x and decrease rate y) according to the assumed distribution can be derived.
[0114]
[Equation 3]

[0115]
According to the parameter generation method for the probabilistic uncertain environment as described above, the user can specify each period from the assumed distribution of each element in the final period only by specifying the period for considering the probabilistic uncertain environment element and capital investment. A parameter indicating the variation rate of each variation of the element can be automatically calculated by back calculation. Therefore, effective parameters can be automatically generated efficiently without the need for the user to generate parameters for the stochastic uncertainty environment.
[0116]
[2-6. Risk assessment standard generation]
In the capital investment risk evaluation system 100 of the present embodiment, as described above, the risk evaluation standard 122 representing the user's risk preference standard is used for evaluation of the decision tree, but the setting of the risk evaluation standard 122 is not easy. Therefore, in the present embodiment, the risk evaluation criterion generation means 109 learns the risk preference of the user by allowing the user to specify which option of the risk evaluation criterion to select.
[0117]
FIG. 23 is a block diagram showing a detailed configuration of the risk evaluation criterion generation means 109 for learning the user's risk appetite. As shown in FIG. 23, the risk evaluation criterion generation unit 109 includes an information presentation unit 2301, a decision making unit 2302, a user selection result acquisition unit 2303, and a learning unit 2304.
[0118]
With this configuration, the risk evaluation criterion generation unit 109 selects an option based on the magnitude of the risk-adjusted value calculated based on the given risk evaluation criterion 122 in step S1004 (FIG. 10) of the decision tree evaluation procedure. The information presenting means 2301 presents the risk and return of each option to the user, and the user selection result obtaining means 2303 obtains the user's selection result. Information regarding the option selected by the user is notified to the decision tree evaluation unit 103 by the decision making unit 2302 as a decision result 2311.
[0119]
In this way, by making a decision based on the user's selection result, the user's risk appetite can be recognized indirectly and partially, so the partial risk appetite that can be derived from the user's selection result is learned means 2304. To learn and accumulate as evaluation rules 2312. For example, as shown in FIG. 24, when the risk evaluation criterion is expressed by a function f, a constraint expression related to f can be generated from one selection result of the user. Specifically, when the user determines that the value of option 2 is higher than that of option 1 in FIG. 24, the constraint expression “risk adjusted value of option 1 <risk adjusted value of option 1” can be learned. By repeating this process, the risk evaluation reference 122 can be generated from the evaluation rule 2312 that is a set of constraint expressions as a result of learning the user's risk appetite.
[0120]
Specifically, f can be estimated from the accumulated set of constraint equations by mathematical programming and numerical solution. Which type of mathematical programming can be applied depends on the function form of f, but a function system that satisfies the constraint is selected, and the parameters of the function are solved by mathematical programming. If it cannot be solved, it can be approximated by a numerical solution. In addition, there is a case where it is possible to select an option based on the evaluation rules 2312 accumulated in the past without inquiring the user. In this case, acquisition of the user selection result by the user determination unit 2303 can be omitted, and the decision making unit 2302 can automatically make a decision based on the evaluation rule 2312.
[0121]
According to the risk evaluation standard generation method as described above, the risk evaluation standard can be automatically generated only by the user selecting the risk evaluation standard with reference to the information on the options of the risk evaluation standard. Therefore, it is possible to efficiently and automatically generate a risk evaluation standard that reflects the user's risk appetite without the need for the user to generate a risk evaluation standard.
[0122]
[3. Other Embodiments]
Note that the present invention is not limited to the above-described embodiments, and various other forms can be implemented within the scope of the present invention. First, the system configuration, the configuration of each part, the processing procedure, the data structure, the display format, etc. shown in each figure are merely examples, and the specific configuration, processing procedure, data structure, display format, etc. can be freely selected. . Furthermore, as described above, the present invention is suitable for making decisions on capital investments for relatively long periods of time, such as steel, chemicals, and power plants. The present invention can be widely applied in the case of performing, and in either case, excellent effects can be obtained similarly.
[0123]
【The invention's effect】
As described above, according to the present invention, by generating and evaluating a decision tree by dividing a scenario representing the external environment into a probabilistic and a non-probabilistic one, various external situations can be rationalized and To provide a capital investment risk evaluation method and system that can be set efficiently and that can make rational and efficient decisions on capital investment while comprehensively considering evaluation results in various external situations. be able to.
[Brief description of the drawings]
FIG. 1 is a block diagram showing the overall configuration of a capital investment risk evaluation system according to a basic embodiment to which the present invention is applied.
FIG. 2 is a flowchart showing an outline of processing for capital investment risk evaluation by the capital investment risk evaluation system shown in FIG. 1;
3 is a data display diagram showing an example of a decision tree generated by the equipment investment risk evaluation system shown in FIG. 1. FIG.
4 is a data display diagram showing an example of a set of macro external scenarios used for generating the decision tree shown in FIG. 3. FIG.
FIG. 5 is a data display diagram showing an example of a binary tree representing a one-period variation in each element of the probabilistic uncertainty environment as the probabilistic uncertainty environment used for generating the decision tree shown in FIG. 3;
6 is a case in which a binary tree showing fluctuations of each element of the stochastic uncertainty environment shown in FIG. 5 is combined over a plurality of periods as the stochastic uncertainty environment used for generating the decision tree shown in FIG. The data display figure showing an example.
7 is a data display diagram showing an example of a strategy option used for generating the decision tree shown in FIG. 3. FIG.
FIG. 8 is a flowchart showing a subroutine of a decision tree generation procedure shown in FIG.
9 is a graph showing an example of risk evaluation criteria used in the decision tree evaluation shown in FIG.
10 is a flowchart showing a subroutine of a decision tree evaluation procedure shown in FIG.
FIG. 11 is a flowchart showing a subroutine of a return risk calculation procedure performed in the decision tree evaluation procedure shown in FIG. 10;
12 is a data structure diagram showing a data structure of an evaluation result obtained by the decision tree evaluation procedure shown in FIG.
13 is a block diagram showing a detailed configuration of an evaluation result analysis display unit shown in FIG. 1. FIG.
FIG. 14 is a flowchart showing a subroutine of the display result analysis display procedure shown in FIG. 2;
15 is a data display diagram showing a display example in a three-dimensional table format with each element as each dimension as an example of evaluation result set display in the display result analysis display procedure shown in FIG. 14;
16 is a data display diagram showing an example of individual evaluation result display when a plurality of evaluation results are designated in the display result analysis display procedure shown in FIG.
FIG. 17 is a data display diagram showing an example of simulation result display in the display result analysis display procedure shown in FIG. 14;
18 is a block diagram showing a detailed configuration of a non-stochastic uncertain environment generation unit and an evaluation result storage unit shown in FIG.
FIG. 19 is an explanatory diagram showing an example of creating a macro external scenario by the non-stochastic uncertain environment generation unit shown in FIG. 18;
20 is a data display diagram showing an example of an allowable risk specification used for evaluation result evaluation by the evaluation result storage unit shown in FIG.
FIG. 21 is a conceptual diagram showing a scenario space that does not need to be confirmed and a scenario space that needs to be confirmed in the case of performing non-stochastic uncertainty environment generation by feeding back the test results by the evaluation result storage unit shown in FIG.
22 is a data display diagram showing assumed distribution information used for parameter generation of the stochastic uncertainty environment by the parameter generation unit shown in FIG. 1;
FIG. 23 is a block diagram showing a detailed configuration of a risk evaluation criterion generation unit shown in FIG. 1;
24 is an explanatory diagram showing an example of option presentation by the risk evaluation criterion generation unit shown in FIG. 23 and accumulation of evaluation rules based on a user selection result.
[Explanation of symbols]
100 ... Capital investment risk assessment system
101 ... Evaluation case setting means
102. Decision tree generation means
103 ... Decision tree evaluation means
104 ... Evaluation result storage means
105 ... Evaluation result analysis display means
106: Cash flow calculation means
107: Non-stochastic uncertain environment generation means
108: Parameter generation means
109 ... Risk assessment standard generation means
111 ... Non-stochastic uncertainty environment (macro external scenario)
112 ... Stochastic uncertainty environment
113 ... Strategic options
121 ... Decision tree
122 ... Risk assessment criteria
123 ... Evaluation result
301: Decision node
302 ... External status designation node
303 ... External probability selection node
304 ... Scenario
310 ... start node
311 ... Decision-making branch
312 ... External situation designation branch
313 ... External probability selection branch
1201 ... Evaluation case
1202 ... Result information
1211 ... Macro external scenario
1212 ... Strategic options
1213 ... Risk assessment criteria
1221 ... Expected value of discounted present value
1222 ... Distribution present value distribution
1223 ... Decision result information
1301 ... Display condition setting means
1302 ... Evaluation result set display means
1303 ... Individual evaluation result display means
1304 ... Simulation means
1305: Simulation result display means
1306: Comprehensive evaluation determining means
1801 ... Scenario element of non-stochastic uncertainty environment
1802 ... Basic scenario
1803 ... Macro external scenario generation means
1811 ... Acceptable risk specification
1812: Inspection means
1813 ... Accumulation means
2301 ... Information presentation means
2302 ... Decision making means
2303 ... User selection result acquisition means
2304: Learning means
2311 ... Decision result
2312 ... Evaluation rules

Claims (10)

In a capital investment risk evaluation method that uses a computer having an input / output unit, a storage unit, and an arithmetic processing unit to evaluate a capital investment plan in consideration of risks in an uncertain external environment,
A non-stochastic uncertainty environment that represents the uncertainty of a non-stochastic external environment that cannot be statistically modeled as a stochastic process, or a stochastic that represents an uncertainty of a stochastic external environment that can be statistically modeled as a stochastic process If you define an uncertain environment, a strategy option that represents the option that the user wants to evaluate, and a risk assessment standard that represents the user's risk appetite criteria,
Non-probabilistic uncertain environment generation means realized by a computer processing unit and a program for controlling the computer, decision tree generation means, decision tree evaluation means, evaluation result storage means,
Based on the scenario element instruction from the user input by the input / output unit, the non-stochastic uncertain environment generating means determines a set of scenario elements of the non-stochastic uncertain environment, and the scenario element is determined from the set. A non-stochastic uncertain environment generation step that combines to generate a non-stochastic uncertain environment scenario;
Stochastic uncertainty environment and strategy option given as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data, and the non-stochastic uncertainty environment generation step A decision tree generation step of generating a decision tree by the decision tree generation means using the non-stochastic uncertainty environment scenario generated in
Based on the risk evaluation criteria given as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data, the decision tree evaluation unit performs the decision tree determination step A decision tree evaluation step for evaluating the decision tree generated in
An evaluation result accumulation step of storing the evaluation result of the decision tree in the decision tree evaluation step in the storage unit by the evaluation result accumulation unit;
The decision tree generation step includes a decision node that generates a decision branch corresponding to each option of the strategy option, and an external situation specification node that generates an external situation specification branch corresponding to each scenario of the non-stochastic uncertainty environment Generating a decision tree using three types of nodes: and an external probability selection node that generates an external probability selection branch corresponding to each variation of each element of the probabilistic uncertainty environment;
The decision tree evaluation step calculates a return and a risk for each decision branch corresponding to each option of the strategy option for the decision tree generated in the decision tree determination step, and calculates the risk from the calculated return and risk. A method for evaluating capital investment risk , comprising calculating a risk-adjusted value based on an evaluation criterion and selecting a strategic option option that maximizes the risk-adjusted value . The evaluation result accumulation step includes:
  Based on the permissible risk specification set in advance as a condition for redundant and unconfirmed evaluation results, it is checked whether the evaluation result of the decision tree in the decision tree evaluation step is redundant and does not require confirmation. An inspection step to eliminate the results;
  Including an accumulation step for storing evaluation results remaining without being excluded by this inspection step;
  The non-stochastic uncertain environment generation step includes a step of generating a non-stochastic uncertain environment scenario by selecting a combination of scenario elements based on the information obtained by the inspection step.
The capital investment risk evaluation method according to claim 1. Using the evaluation unit setting means and the evaluation result analysis display means realized by the arithmetic processing unit of the computer and a program for controlling it,
  Based on a user instruction given by the input / output unit, the evaluation unit setting means, as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data Given probabilistic uncertainty environment and strategy options, and And an evaluation unit setting step for setting each combination of three types of elements as a non-stochastic uncertainty environment scenario generated in the non-stochastic uncertainty environment generation step as an evaluation unit, and storing it in the storage unit;
  Evaluation result generated by the evaluation result analysis display means for displaying the evaluation result obtained by performing the decision tree generation step and the decision tree evaluation step for each evaluation unit, and displayed by the input / output unit Analysis display step,
The capital investment risk evaluation method according to claim 1 or 2, characterized by comprising: By using a risk evaluation standard generating means realized by the arithmetic processing unit of the computer and a program for controlling it,
A risk evaluation criterion generating step for presenting information on each option of the risk evaluation criterion to the user through the input / output unit by the risk evaluation criterion generating means, and generating a risk evaluation criterion based on a user instruction,
The capital investment risk evaluation method according to any one of claims 1 to 3, wherein: Using parameter generation means realized by an arithmetic processing unit of the computer and a program for controlling it,
When taking into account the variation of each element of the probabilistic uncertainty environment over a period of T (T is a natural number of 1 or more) by this parameter generation means, from the assumed distribution of each element in the T period, A parameter generation step for generating by calculating back a parameter indicating a variation rate of each variation;
The capital investment risk evaluation method according to any one of claims 1 to 4, wherein: The evaluation result analysis display step includes
Image data in a predetermined format expressing the set of evaluation results stored in the storage unit as viewpoints of three types of elements, the non-stochastic uncertainty environment, the strategy option, and the risk evaluation criteria. Generating and displaying by the input / output unit,
The capital investment risk evaluation method according to any one of claims 3 to 5, characterized in that: The predetermined format is a three-dimensional format having the three types of elements as dimensions.
The capital investment risk evaluation method according to claim 6. The evaluation result analysis display step includes
A set display step of generating image data of the predetermined format expressing the set of the evaluation results, and displaying the data by the input / output unit;
An individual display step of generating image data that individually expresses details of evaluation results selected from the set of evaluation results based on a user instruction given by the input / output unit and displaying the image data by the input / output unit When,
Based on a user instruction given by the input / output unit, a simulation is performed using an evaluation result selected from the set of evaluation results to generate image data representing the simulation result, and the input / output unit Including the simulation result display step to display,
The capital investment risk evaluation method according to claim 6 or 7, characterized in that: In a capital investment risk evaluation system that uses computers to evaluate capital investment proposals considering risks in an uncertain external environment,
Non-stochastic uncertainty environment table to the probability uncertainty of stochastic external environment for statistically modeled as a stochastic process which represents the uncertainty of the non-stochastic external environment that can not be statistically modeled as stochastic processes Defined as an uncertain environment, a strategy option that represents the option that the user wants to evaluate, and a risk evaluation criterion that represents the user's risk appetite criteria,
Based on a scenario element instruction from the user input by the input / output unit, a set of scenario elements of a non-stochastic uncertainty environment is determined, and a scenario element of the non-stochastic uncertainty environment is determined by combining the scenario elements from the set. A non-stochastic uncertain environment generating means to generate;
Stochastic uncertainty environment and strategy option given as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data, and the non-stochastic uncertainty environment generation means A decision tree generating means for generating a decision tree using the non-stochastic uncertainty environment scenario generated in
Based on the risk evaluation criteria given as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data, the decision tree generated by the decision tree determination unit A decision tree evaluation means for performing an evaluation;
Evaluation result accumulation means for storing the evaluation result of the decision tree by the decision tree evaluation means in the storage unit;
The decision tree generation means includes a decision node that generates a decision branch corresponding to each option of a strategy option, and an external situation specification node that generates an external situation specification branch corresponding to each scenario of a non-stochastic uncertainty environment. And an external probability selection node that generates an external probability selection branch corresponding to each variation of each element of the probabilistic uncertainty environment, and configured to generate a decision tree using three types of nodes:
The decision tree evaluation means calculates a return and a risk for each decision branch corresponding to each option of the strategy option for the decision tree generated by the decision tree determination means, and calculates the risk from the calculated return and risk. A capital investment risk evaluation system configured to calculate a risk-adjusted value based on an evaluation criterion and to select a strategy option that maximizes the risk-adjusted value . In a capital investment risk evaluation program that evaluates a capital investment plan in consideration of risks in an uncertain external environment using a computer having an input / output unit, a storage unit, and an arithmetic processing unit,
A non-stochastic uncertainty environment that represents the uncertainty of a non-stochastic external environment that cannot be statistically modeled as a stochastic process, or a stochastic that represents an uncertainty of a stochastic external environment that can be statistically modeled as a stochastic process If you define an uncertain environment, a strategy option that represents the option that the user wants to evaluate, and a risk assessment standard that represents the user's risk appetite criteria,
Based on a scenario element instruction from the user input by the input / output unit, a set of scenario elements of a non-stochastic uncertainty environment is determined, and a scenario element of the non-stochastic uncertainty environment is determined by combining the scenario elements from the set. A non-stochastic uncertainty environment generation function to generate,
Stochastic uncertainty environment and strategy option given as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data, and the non-stochastic uncertainty environment generation function A decision tree generation function for generating a decision tree using the non-stochastic uncertainty environment scenario generated in
Based on the risk evaluation criteria given as input data by the input / output unit, data stored in the storage unit, or data generated from any of the data, the decision tree generated by the decision tree determination function A decision tree evaluation function for performing an evaluation;
Causing the computer to realize an evaluation result storage function for storing the evaluation result of the decision tree by the decision tree evaluation function in the storage unit;
The decision tree generation function includes a decision node that generates a decision branch corresponding to each option of a strategy option, and an external situation specification node that generates an external situation specification branch corresponding to each scenario of a non-stochastic uncertainty environment. And a function for generating a decision tree using three types of nodes: an external probability selection node that generates an external probability selection branch corresponding to each variation of each element of the probabilistic uncertainty environment,
Wherein the decision tree evaluation functions, the decision tree generated by the decision tree determination function, the return and risk was calculated for each decision branch corresponding to each option of the strategic option, from the calculated return and risk A program for evaluating capital investment risk , comprising a function of calculating a risk-adjusted value based on a risk evaluation standard and selecting a strategic option option that maximizes the risk-adjusted value .

Images